aie. OF oe a eS ee eT re UE yy Peis (AGO = Sz En Esra Ss gebisbespapwchcee Agta HiG 553 Pentre tr pet hy Sesepaere nye Secres $ PE BUT i AR ates jl ote e; 7 i oF : "I J i 1 n DLR THAEE s ESL tin ANA EaE ipeatt 4 tat 343 wae. ft ae ied 34 ARES, £ x 3 4 f Ot eto ee SER oe al es Soo eey: BALES ih fr ! 5, tevksy i n Aen he 21 Aas a by are Or as RH He ‘ it} Hs Me iy? i} Veeipaays Mi Baht, ated A Meet Uy ronan u ¥ Kin 1] ay ( TODe Any he THE AMERICAN JOURNAL CHARLES R. BARDEEN University of Wisconsin Henry H. DonaLpson The Wistar Institute Simon H. GAGE Cornell University OF EDITORIAL BOARD G. Cart HUBER University of Michigan GEORGE S. HuNTINGTON Columbia University Henry McE. KNower, SECRETARY University of Cincinnati VOLUME 17 1914-1915 ANATOMY FRANKLIN P. MALL Johns Hopkins University J. Puayratr McMourricu University of Toronto GrorGeE A. PIERSOL University of Pennsylvania THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. CONTENTS 1914-1915 No.1. NOVEMBER E. V. Cowpry. The comparative distribution of mitochondria in spinal ganglion cells of vertebrates. Fourteen figures (three plates)......... F. W. Toyne. The anatomy of a 17.8 mm. human embryo. Eight figures Rosert BENNETT Bean. The stature and the eruption of the permanent teeth of American, German-American and Filippino children. Deduc- tions from the measurements and examination of 1445 public school chil- dren in Ann Arbor, Michigan, and 776 in Manila, P. I. Five diagrams No. 2. JANUARY Otto FREDERIC KAMPMEIER. On the origin of lymphatics in Bufo. Thirty- END). TEI CUR ESE 5 te a iehay Brak ea Pes Rane CEP eh me Rie oN ort Henry K. Davis. Astatistical study of the thoracic ductinman. Thirty- Osa OAT Cama PANN fea ff ot arch EM calc ch RE alas La an hebae 6h Meuhatte Stee No. 3. MARCH _Ricwarp E. Scammon. The histogenesis of the selachian liver. Forty-five FS MTONE (SEEN: PIAUER) Pale cs a Ge teres Shel ee iG aula cc Aree, Bia Bek J. A. BADERTSCHER. The development of the thymus in the pig. I. Mor- phogenesis. Twelve figures (two plates).................0ec000- MarGARET REED LEWIS AND WARREN Harmon Lewis. Mitochondria (and other cytoplasmic structures) in tissue cultures. Twenty-six figures. . No. 4. MAY RanpDourx West. The origin and early development of the posterior lymph heartvmetnevenck. Mourteen: figures... 0.0... vee sels we se ee eens oe ke J. A. BaprrtscHer. Development of the thymus in the pig. II. Histo- genesis. -Pimpernlates: (eight figures) ..2.05 550.5000. ease esas sc eee ees L. Bork. On the premature obliteration of sutures in the human skull.... 113 161 211 245 317 339 403 495 ; wee" eesheus, bys ok THE COMPARATIVE DISTRIBUTION OF MITO- CHONDRIA IN SPINAL GANGLION CELLS _ OF VERTEBRATES! E. V. COWDRY Department of Anatomy, Johns Hopkins University FOURTEEN FIGURES (THREE PLATES) CONTENTS TET BI RONG HOMER IIICO HN ae era es cen ey rts te RE a Sc ude c a AER aN 1 TE AHR SIEM HUTS oo ee a Ree ene TERRA Oe RNG SRRTCIE ch cece NCR Bitrate ot Re eee 2 nee tie ea TEC RIMeCEMOMS 1 aches Gio gail, tA MARAE Bick es sheune clste ee wahoo, «th iin Jae ene PSST AUTO TAS ee ree eee a NL rr ane Tu Ae ALPINE ps (UP OMECY Woe) cy onlay ea 11 TDS OTS Vana e be Oe a VN RY 8 a ee, DACA Seta Vani RT NA ce RR Te OEE ne 16 TRESIIILTER, 5 beste thc MMe ay ok es Ue gi pe ERM CORE Toe MIU ould gua aR Og aT a toot ae SOO gd a 50 BEA Grace melanie bre ey.n Teh Se keira, Ee wien iy LR hake yt ibaa eb duty: 50 ID AUCTIONED Koa See ene Re enh, eRe a ge ne 51 ule sernerars lniy! orn epee an eee Rae eke Oi Oy Mews ek GOP ae ee oe 51 LSUGINETE 6 a5 57h 0 REN ee cet ee ee te 52 IME femcerial meine Ne te Ae eo eet) oe De es 52 Mivoler Ge ten ume meer Soe, on ks Nn on Ree ea, A eee oe a aas bh atlanaeee 53 ‘This work has been aided by a grant from the Elizabeth Thompson Science Fund. 31 32 iS) S]LAI TGR CLO) 00 Dee ren RIM eia. Shad mio omiin Ours Ol Coren omit oleh orton on Soe ranels eros O C 54 @ranral Werves s24 2 ciict ah eR ere oe eee Bo NAN eae ae Oke oc rare 54 ININ:. OLLA GE OMILS Se Svea reset Boeri oe te tore, ets sce ar ica ear eae 54 ING: “OD GHCUS ors Aero eecoe R eeLe S a tgete 54 IN|. OCUlOMOLOTIIS# seem rees er teat o Seeclens Fa stake eee eee eee 54 N.S GE OCHIC PISO Aree sate wee eee perc ok ak ws ocatee cs RI Oe ro Se 54 IN. . EEN @ON UMS pacha cela oses) Sen Sar oss a. ois wn eo cle tn genes Gree ES 55 Ne abducenisa eect roe oes sake ects nce otitis er eee 56 INE Pa Cia liste cerns epee tral peers Src Ayigtits s,s Reviaireye: uadesaa seo rok ae Aa 57 ING SA CUS EL CUS eae eever terete tes oer Ghe vie a ws saat otcle Sie a Ri Opa ae 57 Necclossophamymeecusten sot otc atiecae? aoe neo Gee hero ee 58 IBN GI AUIS Giy Rerutp ou ataiog 6 dine eae Ree Rea ee naman PG TERS aa onO otns'o o - 59 INGBAICCESS OTIS PRR Roe frie ass tachi AA.» GlrbIN ISR IAC COTE a re ae ee 63 Neg VOPLOSSUB ts Bees tene © wel Ass so oe tbe wae ee ee oe en 64 Spimallmenyes saree cere aera oe sees ose Sean te ae eee ee 65 ODN) CSHENICE Isis a eho ead ro eee VaR one Reno: oa ties Say ERomp rs alka ov < 65 EMSC LORE aise ee eee ee Ree ates wie 2 Sic va aveletd HSE SiG ere SS EO ve ee 68 Hees Se Me a aa eae ee ae ead ee mR Mereee IN ee ee Weim oS 68 | DEN iia a cas eats crete tiated ae ee eee ae a ee We Rl het dan ox 68 IN oon Ae Racin c Sates ep eC eee een chs Ke tore ne GiGi one mad o0'0 de 69 Vascularisiysthemisner cepcimecscs @. iat se Suck eke ooee Stare ger Rene ae ae 70 15 [EER rR ASS o. Msves As choke okt el Ren ie tortoise at eee, etn SAU ola no 7G SIMUSMVEMOSUS eee cea ake eens kod ae ee re a Red aA EN UNO Gs, ees Ree eRmran eee Ares chon AonB con halen ia aa © 70 IRCA HELO TOIKG! (Wy seo te GARNER ore a's, ie lo Rie o romb'aacut bp,on dion oath. ql NetGear trite ey cls.) vss,1cte ease Ae ACRE eo ee ee ee 71 The hibpvemuniel Gerace cia ae = doves ah eo ee ae er orc ee al (AT LCTICS MOE TRE ORs ae i oe ae RO ORO ee 72 Systemucqderivatives of the aorticsarehest-c.-% - 1. 5. eee eee 72 Palmom ayn anvenies:\..\ 2.5 emcee tele sem eae moe cee bee ee eee 76 Derivatives of the occipital and cervical segmental vessels.......... 76 Parietaluarteries of the tramk.....)..21.042 ences ee ode oe VaASCeral@anvenleStmnciace cae Stan cae eee ee ee eee 79 Were s AHETICS ea Meer ns acs vo clea ARIA Y Slord ete Bie ee ee 82 \WGTO EES oan ested SSRIS Gk ae ee OE ME en EID VT eo os 83 Antenlonand commonicardimal system sqse eee eee eee eee 83 Rostenionicandimal’sy.stemiy...7.¢on otek c on. eee eee 87 ‘PRontalSsyis hemmee cra ae dees ou isaacon vend Haske sia. ease ee aa ROS on ones Pena 90 Pulmonary veins (see left atrium). I Lina) of NEN ACIS eBags eecunnce Ra he to cee ore a REC CRTRAR e piea rao Arar BARA AC cal 3 Gd 91 SEVCOUE) TUNATI RAI SONG, oS occ oy oueen bao b oe sdonoeeEsEebavoeuoLe 91 SACCuUss UTM ATISNMe LOIN, Wet crce Cains sehen oe Oe rae lon Tee 91 SAC CUSUMESEMUCE CU Stet: so ve iar et tai sPecuat aaa ee nea Ree oa 92 @istern ac ly lites ee esse thew leks Ws Waneihe s clelis on SeA ESE CR ee 92 Saccialyamphaticignostenlionesann cata iccs ae. Selec ae ee 92 ANATOMY OF A 17.8 MM. HUMAN EMBRYO 33 INTRODUCTION This work was undertaken at the suggestion of Prof. C. 8. Minot, and carried on for the most part in his laboratory during the years 1906 and 1907 when the author held an Austin Teach- ing Fellowship in Histology and Embryology at the Harvard Medical School. Its completion, unfortunately delayed by other work, has been accomplished in the Anatomical Depart- ment of the University and Bellevue Hospital Medical College. During the progress of the work many helpful suggestions have been received from Professors Minot and IF. T. Lewis of Harvard, and Prof. H. D. Senior of Bellevue, for which I am very grateful. I also desire to express my gratitude to Mr. W. T. Oliver of Lynn, Massachusetts, for the careful manner in which he has reproduced in finished form my original drawings. The reconstructions upon which this work is based, were made from transverse sections of Embryo 839 of the Harvard Embryo- igrical Collection, chiefly by the modified graphic reconstruction inethod of His. The shading usually has been inferred from a study of the sections, but in a few instances wax models were made of regions requiring a fuller interpretation. This embryo (extra-uterine) measured in formalin 17.8 mm., greatest length, with a neck breech of 16.7 mm. The greatest length in 80 per cent alcohol was 13.6 mm. In previous papers in which this embryo was referred to (Thyng ’08, and Lewis and Thyng ’08), the latter measurement was given. EXTERNAL FEATURES The external features of this embryo are seen in profile view in text figure 1, a reproduction of figure 104 in Minot’s (’10) “Laboratory text-book of embryology,” also in part in plates 3 and 5. The neck-bend is approximately a right angle; the cephalic flexure is also very nearly a right angled bend, so that the oral aperture is in close proximity to the cardiac region. The dorsal flexure has disappeared almost entirely, only a. slight elevation persisting to mark its earlier position. Above this THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 1 34 F. W. THYNG Text fig. 1 Human embryo of 17.8 mm. (H. E. C. 839). X 5 diams. (after Minot).? elevation there is a shallow depression, said to disappear in the course of development. A distinct groove, extending transversely between the medial angles of the developing eyes, separates the forehead from the root of the nose. The maxillary process of either side has joined the adjacent lateral and median nasal processes, obliterating the naso-optic grooves. The nares are open, but separated by a rather low, broad septum. 150. Reduced to X 75. REFERENCES l., ventral cephalic lymph sinus j.d., j.8., right and left external jugular 1.l., U.l., lateral extensions or wings of veins the sinus which are joined to it fur- — m.c., mouth cavity ther forward 184 ORIGIN OF LYMPHATICS IN BUFO PLATE 1 OTTO FREDERIC KAMPMEIER Sees Oe iy nets S Veiuts 1 es < zs es RG Das ~ 1100. REFERENCES l., l., sinus anlagen t.j7., jugular tributary (see footnote 5, j.d., j.s., right and left external jugular page 168) veins b.c., blood corpuscle 2and3 Successive sections from the left side. Slide 1, row 4, sections 19 and 20. 4 From the same side, three sections caudad. Slide 1, row 4, section 25. Compare the lymphatic rudiment (1) with the mesenchyme in regard to yolk, and note its solidity. 5 and 6 Two alternate sections from the right side. Slide 1, row 4, sections 18 and 20. Observe the cavity of the lymphatic (1) in the second sketch. 186 ORIGIN OF LYMPHATICS IN BUFO PLATE 2 OTTO FREDERIC KAMPMEIER 187 PLATE 3 EXPLANATION OF FIGURES 7 Transverse section of a 6 mm. toad embryo through the sinistro-ventral region of the head. Series 54, slide 1, row 4, section 8. X 1100. REFERENCES /., sinus anlagen j.s., left external jugular vein Sand9 Transverse sections of another 6 mm. toad embryo through the dextro- ventral region of the head. Series 53, slide 1, row 5, section 13, and row 6, section 4.- 1100. Eight sections intervene between these two levels. REFERENCES (1.1, 1.2, 1.3, sinus anlagen j.d., right external jugular vein t.j7., wall of a tributary cut tangentially ORIGIN OF LYMPHATICS IN BUFO PLATE 3 OTTO FREDERIC KAMPMEIER PLATE 4 EXPLANATION OF FIGURES 10 Transverse section of a 7 mm. toad embryo through the ventral region of the head. Series 52, slide 1, row 4, section 5. 300. 11 A small area near the lower left corner of the same section (fig. 10). x 1100. REFERENCES j.d., j.s., right and left external jugular —_(l., /., simus anlagen veins m.c., mouth cavity 190 4 PLATE ORIGIN OF LYMPHATICS IN BUFO OTTO FREDERIC KAMPMEIER ¢ NS i.) 38 C ee PST Wt Ss... SOA ) ee ees e Bag 0590.0 2% B). |. g Be P06 ae PD? MV ea 8 Seo 10 er eee were RM, Fee 3 : 191 PLATE 5 EXPLANATION OF FIGURES 12 Transverse section of an 8 mm. toad embryo through a little more than the right half of the territory ventral to the mouth cavity (cf. fig. 1). Series 27, slide 1, row 5, section 8. > 500. REFERENCES l. to l., plexiform anlage of sinus e.c., right external carotid artery j.d., right external jugular vein m.a., muscle anlagen 13 Transverse section of a 9 mm. toad embryo through the ventral region of the head. Series 2, slide 1, row 8, section 5. X 300. REFERENCES l., multilocular anlage of sinus j.d., j.8., right and left external jugular t., s., tissue strands and trabeculae, veins remnants of walls between originally — m.c., mouth cavity independent channels 192 ORIGIN OF LYMPHATICS IN BUFO PLATE 53 OTTO FREDERIC KAMPMEIER 2. ti or ae tee Zia Gee a. 0 —e 1100. REFERENCES /., lumen of sinus d., cellular débris, probably vestiges t., s., tissue spurs and trabeculae of former partitions m., mesenchymal cell b 194 ORIGIN OF LYMPHATICS IN BUFO OTTO FREDERIC KAMPMUIER 195 PLATE 6 14 PLATE 7 EXPLANATION OF FIGURES 15 Transverse section of a 9 mm. toad embryo through the middle of the trunk. Series 2, slide 3, row 7, section 2. X 150. REFERENCES l.s., l.d., superior and inferior lateral p.m., p.l., medial and lateral divisions lymphatic ducts of the postcardinal vein i.v., intersegmental vein w., Wolffian or pronephric duct m.s., mauscle segment 16 Asmall area of a section of a7 mm. toad embryo through the left mid trunk region. Series 52, slide 3, row 4, section 16. X 1100. REFERENCES l., anterior tip of a blind anlage of the 7.v., intersegmental vein developing superior lateral lymph — ep., epidermis duct m.s., myotome PLATE 7 ORIGIN OF LYMPHATICS IN BUFO OTTO FREDERIC KAMPMEIER 15 197 PLATE 8 EXPLANATION OF FIGURES 17, 18, 19, 20, 21 and 22 Transverse sections to show different levels of the lymphatic anlage the tip of which is pictured in figure 16. Series 52, slide 3, row 4, .section 24; row 5, sections 2, 6, 9, 11; row 6, section 1. X 1100. REFERENCES l., anlage of the superior lateral lymph 7.v., intersegmental vein duct ep., epidermis m.s., myotome 198 ORIGIN OF LYMPHATICS IN BUFO . PLATE 8 OTTO FREDERIC KAMPMEIDR 199 PLATE 9 EXPLANATION OF FIGURES 23, 24 and 25. Transverse sections to show three additional levels of the lym- phatic rudiment illustrated in the sketches on plate 8. Series 52, slide 3, row 6, sections 4, 11, 18. X 1100. 26 Transverse section of a 6 mm. toad embryo through approximately the same region (cf. figs. 17-25) to show an earlier condition, evidently, of the superior lateral lymph duct. Series 53, slide 4, row 5, section 17. X 1100. REFERENCES l., anlage of superior lateral lymph duct — ep., epidermis 7.v., intersegmental vein m.s., muscle segment 27 Transverse section of a6 mm. toad embryo at the level of the left anterior lymph heart. Series 53, slide 3, row 3, section 138. X 1100. REFERENCES l.h., posterior portion of the anterior p.m., p.l., medial and lateral divisions lymph heart of the postcardinal l., anterior extent of the developing b.c., blood corpuscles inferior lateral lymph duct c.c., coelomic cavity 7.v., intersegmental vein ep., epidermis m.s., muscle segment 200 9 4 PLATE ORIGIN OF LYMPHATICS IN BUFO OTTO FREDERIC KAMPMEIER 201 PLATE 10 EXPLANATION OF FIGURES 28 Transverse section a short distance behind the left lymph heart indicated in figure 27. Series 53, slide 3, row 4, section 3. X 1100. REFERENCES l., extreme posterior tip of the anterior p.a., dorsal anastomosis connecting anlage of the inferior lateral lymph the two postcardinal divisions duct, shown in figure 27 as a ventral b.c., blood corpuscles sprout of the lymph heart ep., epidermis p.l., lateral division of the postcardinal —w., pronephric duct REFERENCES 29 HKighteen sections caudad of the level represented by figure 28. X 1100. Series 53, slide 3, row 5, section 4. X 1100. m., mesenchymal cells; no indication p.a., ventral postcardinal anastomosis; of a lymphatic anlage other explanations the same as above 202 ORIGIN OF LYMPHATICS IN BUFO PLATE 10 OTTO FREDERIC KAMPMEIER PLATE 11 EXPLANATION OF FIGURES 30 and 31 Successive sections after an interval of nine sections behind that pictured in figure 29. Series 53, slide 3, row 5, section 12. X 1100. REFERENCES /., discontinuous anlage of the inferior p.a., dorsal anastomosis between post- lateral lymph duct; figure 30 shows cardinal divisions the anterior tip w., pronephrie duct p-m., p.l., medial and lateral post- ep., epidermis -ardinal divisions 204 ORIGIN OF LYMPHATICS IN BUFO PLATE 11 OTTO FREDERIC KAMPMEIER 205 PLATE 12 EXPLANATION OF FIGURES 32 and 33 Consecutive sections after an interval of four sections posterior to those represented on plate 12. Series 53, slide 3, row 5, section 16; row 6, section 1. 1100: REFERENCES l., discontinuous anlage of the inferior w., pronephric duct lateral lymph duct. Figure 33 shows — ep., epidermis the extreme posterior tip p.a., dorsal anastomosis between post- p.l., lateral posteardinal division cardinal divisions c.c., coelomic cavity 206 ORIGIN OF LYMPHATICS IN BUFO PLATE 12 OTTO FREDERIC KAMPMEIER PLATE 13 EXPLANATION OF FIGURES 34and 35 Transverse sections of another 6 mm. toad embryo through the right side of the mid trunk region. Series 54, slide 3, row 3, sections 4and10. X 1100. REFERENCES l., discontinuous anlage of the right p.a., dorsal postcardinal anastomosis inferior lateral lymph duct. The b.c., blood cells last figure shows the anlage clinging to the venous wall p.m., p.l., medial and lateral post- cardinal divisions ep., epidermis w., pronephric duct ORIGIN OF LYMPHATICS IN BUFO PLATE 13 OTTO FREDERIC KAMPMEIER 209 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, NO. 2 AS STAI SLiCAte SUD Y Ob THE THORAGKE DUCH IN MAN! HENRY K. DAVIS From the Anatomical Laboratory of the Cornell University Medical College, Ithaca, New York THIRTY-TWO-FIGURES INTRODUCTION It has long been known that the thoracic duct in man presents a certain amount of variation and in addition to the usually deseribed ducts various marked anomalous conditions have been noted. This investigation was undertaken to determine the percentage of occurrence of the different variations of the thoracic duct. An attempt has also been made to explain these vari- ations from an embryological standpoint. The various types of duct that might develop from the primitive embryological network have been indicated and the ducts here described to- gether with those described by other investigators have been divided into corresponding groups. MATERIAL AND METHODS This paper is based upon the records of the dissection of 22 cadavers in the Anatomical Laboratory of the Cornell Univer- sity Medical College, Ithaca, New York. Forty-two cadavers, on which autopsies had been performed, were examined, but many of them had to be discarded on account of injury to the duct at the post mortem. In 11 of these however, the duct was found complete and records were taken of the ducts in these bodies. The other 11 records were taken from bodies which were dissected by the medical students. During the course of ‘From a thesis presented to the faculty of the Graduate School of Cornell University for the degree of Master of Arts, June, 1914. 211 le, HENRY K. DAVIS the dissection of the abdomen, thorax, and base of the neck, I supervised the students’ work so that no injury would occur to the duct and made the dissection of the duct myself. The thoracic duct in each case was injected with a carmine gelatin mass (Lee ’05). At first, I attempted to make the in- jection from the cephalic portion of the duct but could not make the injection mass flow caudad on account of the valves. By experimenting, I found that by exposing the duct Just cephalic to where it pierces the diaphragm and making the injection from this point, that the injecting mass flowed freely caudad as well as cephalad. This seems to indicate that the valves are much more efficient in the cephalic than in the caudal portion of the duct. Before the injection, the innominate, vertebral, subclavian, and internal jugular veins were clamped. This insured a good filling of the duct. After the injecting mass had been allowed to cool and gelatinize, a careful dissection and a natural sized drawing of the duet was made. EMBRYOLOGY According to Sabin (’09) the thoracic duct in human embryos begins in the abdominal cavity at the cisterna chyli as two ducts. These pass cephalad through the thorax, the right duct crossing, at about the level of the 4th thoracic vertebra, dorsal to the aorta to join its fellow of the opposite side. There is thus formed on the left side a single trunk that connects with the jugular portion of the thoracic duct. The jugular portion of the thoracic duct is a caudal outgrowth of the jugular lymph sac on the left side. Sabin, however, has not been able to find the jugular portion, that is, the caudal outgrowth of the right jugular lymph sac, connecting with the thoracic duct. She was able to trace it to the root of the lung but could find no connection of it with the thoracic duct in this region. For this reason the duct as described by Sabin is not a complete bilaterally symmetrical duct. In the embryo there is a distinct right and left duct. These two ducts are connected by numerous cross anastomoses thus forming a plexus of lymphatic vessels along the course THORACIC DUCT IN MAN 2S of the aorta. Sabin (09) reports the first appearance of the cisterna chyli in 23 mm. human embryos. Here it is a definite sac opposite the 3rd and 4th lumbar vertebrae. The thoracic duct is first found in human embryos of 24 mm. In human embryos of 30 mm. the thoracic duct is complete. Sabin (02) in her study of the thoracic duct in pig embryos found it to be essentially the same as she found in human embryos. Lewis (06) found the thoracic duct in the rabbit embryo to be practically the same as Sabin found in the human and pig embryos. Huntington (11) in his study of the lymphatic vessels of the cat states that the thoracic duct is ‘“‘potentially bilaterally symmetri- cal’ and he pictures a bilaterally symmetrical duct in figure 29, plate 22. It is interesting also to note that Sala (99-00) and Pensa (O8—’09) picture bilaterally symmetrical thoracic ducts in birds DIVISION OF THE THORACIC DUCTS INTO GROUPS Assuming that the embryonic thoracic duct is bilaterally symmetrical and that the duct in the adult is produced by the persistence and growth of a part of the embryonic duct and the disappearance of other parts, one might expect to find variations in the adult thoracic duct. These variations depend then upon which portions of the embryonic thoracic ducts atrophy and dis- appear and which continue to develop. These possible varieties of the thoracic duct may be divided into the following types. Type 1. To this type of thoracic duct belong those ducts which would retain more or less the early embryological conditions and would consist of a completely bilaterally symmetrical duct connected by numerous cross anastomoses (fig. 1). Type 2. In this type of thoracie duct we would have caudad the persistence of the original double thoracic duct of the embryo. There would be a right and left duct, which starting in the ab- dominal cavity would pass cephalad through the thorax and at about the level of the 4th thoracic vertebra, the right duct would cross by persistence of one of the embryonic cross anastomosing 214 HENRY K. DAVIS branches to join the duct of the left side. A single trunk would be thus formed which would empty into the venous system of the left side. The cephalic portion of the right duct would fail to connect with the thoracic duct and would remain as the right lymphatic duct (fig. 2). Type 3. In this type of thoracic duct we would have caudad the persistence of the original double thoracic duct of the embryo. There would be a right and a left duct, which starting in the abdominal cavity, would pass cephalad through the thorax and the left duet would cross by persistence of one of the embryonic cross anastomosing branches to join the duct of the right side. A single trunk would thus be formed which would empty into the venous system of the right side. The cephalic portion of the left duct would fail to connect with the thoracic duct and would remain as a left lymphatic duct which would be comparable to the usual right lymphatic duct (fig. 3). Type 4. In this type of thoracic duct, we would have cephalad the persistence of the original double thoracic duct of the embryo. There would be complete atrophy of the caudal portion of the left duet and the cephalic portion of the left duct would join the right duct through the persistence of one of the embryonic cross anastomosing branches (fig. 4). Type 5. In this type of thoracic duct, we would have cephalad the persistence of the original double thoracic duct of the embryo. The caudal portion of the right duct would be completely atro- phied and the cephalic portion of the right duct would join the left duct through the persistence of one of the embryonic cross anastomosing branches (fig. 5). Type 6. In this type of thoracic duct, we would have the persistence of the cephalic portion of the left duct and the caudal portion of the right duct. These two segments would be joined together by the persistence of one of the embryonic cross anas- tomosing branches. The caudal portion of the left duct would be completely atrophied and the cephalic portion of the right duct would persist as the right lymphatic duct (fig. 6). Type 7. In this type of thoracic duct, we would have the persistence of the cephalic portion of the right duct and the THORACIC Right Duct ( Left Duct Right Duct ||/) ON is C/ Si) Right Duct Left Duct Right Duct Left Duct DUCT IN MAN 2|Left Duct Right Duct) | Fig. A Schematic representation of the embryonic lymph channels. Fig. 1 Type 1; schematic representation of the embryonic lymph channels which might persist. Fig. 2 Type 2 which might persist. Fig. 3 Type 3; schematic representation of the embryonic lymph channels which might persist. Fig. 4 Type 4; schematic representation of the embryonic lymph channels which might persist. ; schematic representation of the embryonic lymph channels DAG HENRY K. DAVIS caudal portion of the left duct. These two segments would be joined together through the persistence of one of the embryonic cross anastomosing branches. The caudal portion of the right duct would be completely atrophied and the cephalic portion of the left duet would not connect with the thoracic duct and it would persist as a left lymphatic duct (fig. 7). Type 8. In this type of thoracic duct, we would have the complete persistence of the right embryonic duct. The caudal portion of the left duct would be completely atrophied and the cephalic portion of the left duct would not be connected with the thoracic duct and would persist only as the left lymphatic duct which would be comparable to the usual right lymphatic duct (fig. 8). Type 9. In this type of thoracic duct, we would have the complete persistence of the left embryonic duct. The caudal portion of the right duct would be completely atrophied and the cephalic portion of the right duct would not connect with the thoracic duct and would persist only as the right lymphatic duct (fig. 9). It should be noted that Types 2 and 3, 4 and 5, 6 and 7, 8 and 9 are respectively the reverse of one another in that those channels which persist in one atrophy in the other and vice versa. Group I Winslow (66), Cruickshank (790), Sémmering (’92), and Hommel (’37) deseribe bilaterally symmetrical thoracic ducts. The thoracic ducts start in the abdominal cavity as two ducts which pass cephalad through the thorax, one opening into the venous system of the left side and the other into the venous system of the right side. The right duct lies to the right of the aorta and the left duct on the left side of the aorta. These two ducts are joined together by numerous cross anastomoses. I found no ducts of this type. It is clearly evident that the thoracic ducts described by the above investigators belong to Type 1 (fig. 1). Vig which THORACIC \( MW Left Duct X i OS Ve A \(\| 1. 5 C IN If) D 2 5 Right Duct J K| y \V ‘x! > t /) INN'S | sen OO ATi X / iN OM LY VA OG = aX Ss = Right Duct 8 .5 Type 5; schematic might persist. Fig.6 Type 6; schematic which might persist. Fig. 7 Type 7; schematic which might persist. Fig. 8 Type 8; schematic which might persist. Fig.9 Type 9; schematic which might persist. esa" } representation representation representation representation representation DUCT IN MAN ~ S of the embryoni of the embryonic of the embryonic g » lymph » lymph lymph of the embryonic lymph of the embryonic » lymph Left Duct Left Duct channels channels channels channels channels 218 HENRY K. DAVIS Group II The thoracic ducts of this group (cases 1-6, figs. 10-15) begin in the abdominal cavity as two ducts which extend cephalad through the thorax. The right duct lies to the right of the aorta and the left duct to the left of the aorta. The right duct crosses in the thorax at the level of the 4th thoracic vertebra dorsal to the aorta to join the left duct forming a single trunk which empties into the venous system on the left side at the base of the neck. The two ducts are connected by cross anastomosing channels. This type of duct occurred in 6 cases out of 22, or in 27.27 per cent. This form of duct corresponds to the thoracic duet represented in Type 2. In proportion to the completeness of the persistence of embryonic conditions these ducts have been divided into three divisions, A, B and C. Lower (’80) and Nuhn (’49) describe a similar thoracic duct. Division A. The thoracic duct of this division (fig. 10) begins in the abdominal cavity as two ducts which pass cephalad through the thorax. The right duct at the upper level of the 5th thoracic vertebra begins to incline to the left and passing dorsal to the aorta reaches the left side and at the level of the lower third of the body of the 2nd thoracic vertebra joins the left duct forming a single trunk which continues cephalad to open into the left subclavian vein a short distance from its junction with the in- ternal jugular vein. The right duct lies to the right of the aorta and is situated between the aorta and the vena azygos major. The left duct lies to the left of the aorta. The two ducts are of equal calibre and are connected by numerous cross anas- tomoses. Three of these are especially well developed (1) a cephalic one which connects the cephalic end of the right duct with the left; (2) one which is situated opposite the body of the 9th thoracic vertebra nearly transverse in direction; (3) and one starting in the abdominal cavity on the right side and passing cephalad and to the left to join the left duct opposite the body of the 11th thoracic vertebra. There is no cisterna chyli. It is represented in this case by a plexus of lymphatic vessels. This type of duct occurred in 1 case out of 22, or in 4.545 per cent. THORACIC DUCT IN MAN 219 Internal Jugular Left Internal Jugular Vein ey Subclavian Internal ~ —, fol Trunk Jugular Trunk~_ Subclavian Left Internal Jugular pe |) Left Vertebral Left\, 1st Thoracic Vertebra ZEN ak ‘Subclavian 2a Z : Vein Left 7 eS C= Subclavian” —\\ ae i=» \ Left Vein FS pi int —s Thoracic Innominate tere Duct “~~ Vein e. Innominates ~ Vein Sys me ZO 77 Thoracic WA a = Duct \Wa iS i) 3 laterals 4 \ NS )Collaterals mir. TEs _ Intestinal Trunk SV Ya \ Ny 4 | oe ‘ Lumbar beth Y mphatics Right Left ars el Lumbar Lumbar ANN 4 Lymphatics sil S Lymphatics Mt == << $__— /f lO Fig. 10 Type 1; thoracic duct in a male white subject, age 35. Note the double duct and the abundant cross anastomoses. Fig. 11 Type 1; thoracic duct in a male white subject, age 51. Note the double duct and the reduction in size of the left duct. 220 HENRY K. DAVIS Division B. The thoracic duct of this division (fig. 11) begins in the abdominal cavity as two ducts which pass cephalad through the thorax. The right duct at the upper level of the 5th thoracic vertebra begins to incline to the left and passing dorsal to the aorta reaches the left side and joins the left duct at the level of the lower third of the body of the 2nd thoracic vertebra forming a single trunk which passes cephalad. This trunk at the lower level of the 7th cervical vertebra divides into three branches which do not become united again before opening into the venous system. The most cephalic branch opens into the left subclavian vein a short distance distal to its junction with the left internal jugular vein. The intermediate branch opens into the angulus venosus formed by the junction of the left internal jugular and left subclavian veins and the most caudal branch opens into the left vertebral vein a short distance medial to its Junction with the left innominate vein. The right duct hes to the right of the aorta and is placed between the aorta and the vena azygos major. The left duct lies to the left of the aorta. The two ducts are of unequal size, the right duct being of much greater caliber than the left duct. The cross anastomosing channels in this case are not as numerous as in case | (fig. 10). The chief anastomosis is the cephalic one which joins the cephalic end of the right duct with the left duct. The reduction in caliber of the left duct and the decrease in the number of cross anastomoses point to a stage in the atrophy of the left duct. There is no cisterna chyli present. It is represented by a plexus of lymphatic vessels. This type of duct occurred in 1 case out of 22, or in 4.545 per cent (fig. 2). Division C. In cases 3 and 4 (figs. 12-13), in cases 5 and 6 (figs. 14-15) there is a partial doubling of the caudal portion of the thoracic duct. In each case the caudal portion of the right duct is complete but the left duct has partially atrophied. In case 3 (fig. 12) the thoracic duct begins in the abdominal cavity as two ducts which pass cephalad into the thorax. The right duct lies to the right of the aorta and is placed between the aorta and vena azygos major. It begins to incline to the left opposite the inferior level of the body of the 4th thoracic Internal Left Internal Jugular Vein /Jugular Trunk , . Internal \Subclavian Jugular Trunk Subclavian WA Trunk Trunk ~ So Left Interna! Jugular Vein WN Rett e (7 Ait Cee J Ist Thoracic Vertebra a | WV. ia Us \YSubclavian Left a a ip Subclavian Ve (8 V3 Vein = ( “5 innominate leita < > eS » _S P Innominate: _ Lymph Glands Fi HG V ee Sax ein Ye Ze) S by, ge ~— Thoracic Thoracic Duct ee 4 Collaterals TO ser i = Lumbar [ Lumbar esa | Lymphatics ed \ Lymphatics (WH WS IN: vty Intestina Trunk eas Fig. 12 Type 1; thoracic duct in a male white subject, age 72. Note the incomplete duct of the left side. Fig. 13 Type 1; thoracic duct in a female white subject, age 69. Note the incomplete duct of the left side and the abundant cross anastomoses between it and the duct on the right side. y) 9 9 1 aa Dee, HENRY K. DAVIS vertebra and crossing dorsal to the aorta reaches the left side where it continues its course cephalad to open into the angulus venosus formed by the junction of the left internal jugular and left subclavian veins. The right duct lies to the right of the aorta. It passes up into the thorax from the abdominal cavity and ends at the lower level of the 7th thoracic vertebra. The portion cephalad of this has atrophied. The two ducts are joined together by numerous cross anastomoses. There is no cisterna chyli present. It is represented by a plexus of lymphatic vessels. In figures 13, 14 and 15, the right duct corresponds to the right duct described in case 3 (fig. 12) with the exception that it empties into the left subclavian vein instead of the angulus venosus. The left ducts are essentially the same as in case 3 (fig. 12). In case 4 (fig. 13) there is no cisterna chyli but there is a lymphatic plexus. n case 5 (fig. 14) there are two cysternae chyli. The right duct is a direct continuation of the right cis- terna chyli. The left cisterna chyli is connected to both the right and left ducts. The left duct, however, is not a direct continuation from the left cisterna chyli. In case 6 (fig. 15) there is a single cisterna chyli. The right duct is a direct con- tinuation of this cisterna. The left duct is alsc connected with it. In this case there is a division of the right duct into two branches. This bifurcation takes place at the lower level of the body of the 6th thoracic vertebra and the two branches unite again to form a single trunk at the upper level of the body of the 5th thoracic vertebra. In these 4 cases the lymphatics on the left side of the aorta including the left duct, drain in a caudal direction. Lymph glands are associated with the thoracic duct in cases 3, 4 and 5 (figs. 12-14). A further account of these will be given in dealing with the variations. The arrangement of the two ducts in these 4 cases points to an originally double thoracic duct, as in case 2 (fig. 11). There is represented in these 4 cases another more advanced stage in the atrophy of the left duct. In case 2 there was a reduction in the size of the left duct, while in these 4 cases there is a reduction in size and a complete atrophy of a portion of the left duct. This type of duct occurred in 4 cases out of the 22, or in 18.18 per cent. THORACIC DUCT IN MAN Dae Internal Left Internal Jugular NA asesies Trunk internal 3 poe Trunk Jugular Trunk Subeiavinniicanic Ist ie a Left Interna! Jugular Vein Left \ Ist Thoracic Vertebsa = alt Subclavian Left z ; “LR Aaa & vel Subclavian— 1 > 3) ane Vein Sa ald ; ii. Left ei { < Ad BZA Innominate Lett : ir | Nog P ras 7} i ( \\ \s Lymph = ‘ee : 1K ts Vein Innominateg, alll a If git: Glands “ Pitas } 2 Vein & i fy Ry) fi =o Z i| i} Wy = “Thoracic Duct Collateral Collaterals ollaterals $ nf : ly i volleterals 1 eas Hi $ ia Cisterna {\/) Ef “l) \ | ! i AN) BS Chyli >] sa OES [7 \\\ : Cisterna gekes y VN) = & hyli Nop Evil . Lymph ony 7 f, Intestinal _ (4 Glands Trunk fk N V Cisterna Chyli Intestinal Lymph Cofi PoPe Trunk ¥ s/ Gland eyZ LIN oe CY x VA = A y) | Sh Sexes) f D Rignt 4. my i Left \ |-—/ Lumbar Trunk f 1} 3) Right aaa j 7 De eer Ds Gi ; i umbar Trunk Lumbar. i Yt NS Lumbar = ve Lymphatios Qf ie His toll i Fig. 14 Type 1; thoracic duct in male white subject, age 56. Note the two cisterna chyli and the incomplete duct on the left side. Fig. 15 Type 1; thoracic duct in a male white subject, age 67. Note the incomplete duct on the left side. 224 HENRY K. DAVIS Group IIT Breschet (736) describes a thoracic duct as seen by Haller which was double in its caudal portion. The thoracic duct in this case start in the abdominal cavity as two ducts, which pass cephalad into the thorax, one lying on each side of the aorta. In the cephalic portion of the thorax, the left duct crosses over to the right side and the right and left ducts both open into the left angulus venosus. This form of duct belongs to Type 3 (fig. 3). I found no ducts of this type ainong my own eases and could find no other cases described in the literature. Group IV Butler (03), Lauth (35), Patruban (’44), Diemerbroeck (’85), Cousin (98), and Walther (described by Haller ’46), describe a thoracic duct which starts in the abdominal cavity as a single trunk and passing cephalad into the thorax on the right side of the aorta divides into two branches at about the level of the 4th thoracic vertebra. The right branch opens into the angulus venosus on the right side and the left branch passing dorsal to the aorta opens into the angulus venosus of the left side. The type of duct described by these investigators belongs clearly to Type 4 (fig. 4). I found no ducts of this type among my own cases. Group V The thoracic duct in this type starts in the abdominal cavity as a single duct and passing cephalad into the thorax on the left side of the aorta divides into two branches. The right branch opens into the right angulus venosus and the left branch into the left angulus venosus. I have been unable to find ducts of this type described in literature and there were none among my own cases. THORACIC DUCT IN MAN 225 Group VI In 14 instances in my series (figs. 16-29) the thoracic duct begins in the abdominal cavity as a single trunk which passes cephalad into the thorax and at the level of the 5th to the 3rd thoracic vertebra begins to incline to the left and finally passes to the left of the median line of the bodies of the thoracic vertebrae. The duct continues cephalad and at the level of the 2nd thoracic to the 6th cervical vertebra changes its course passing cephalad, ventrad, and to the left, and then caudad and slightly ventrad to open into the venous system at the base of the neck. fneSisinstances (figs. 16, Li, -20i21 0 23,. 26, 27, 29), the thoracic duct begins to incline to the left opposite the body of the 5th thoracic vertebra; opposite the body of the 6th thoracic vertebra in 1 instance (fig. 28); opposite the body of the 4th thoracic vertebra in 4 instances (figs. 18, 19, 24, 25); and opposite the body of the 3rd thoracic vertebra in 1 instance (fig. 22). The terminal portion of the thoracic duct changes its course opposite the 2nd thoracic vertebra in 1 instance (fig. 20); opposite the 1st thoracic vertebra in 6 instances (figs. 16, 17, 19, 21, 24, 26); opposite the 7th cervical vertebra in 6 instances (figs. 18, 23, 25, 27, 28, 29); and opposite the 6th cervical verte- bra in | instance (fig. 22). The mode of termination is somewhat variable. In 5 in- stances (figs. 16, 20, 22, 27, 28) the thoracic duct terminates by a single opening into the left subclavian vein; into the left subclavian vein by 2 branches in 1 instance (fig. 17); into the left angulus venosus by a single branch in 5 instances (figs. 18, 19, 21, 24, 25); inte the left angulus venosus by 2 branches in 1 instance (fig. 29;) into the left internal jugular by a single branch in 1 instance (fig. 26); and into the posterior wall of the left innominate vein by a single branch in 1 instance (fig. 23). The thoracic ducts begin in the abdominal cavity by the confluence of the lumbar lymphatics and sometimes the intestinal trunk or by a lymphatic plexus in which the lumbar trunks are THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 2 226 HENRY K. DAVIS indistinct. Cases 14 and 18 represent such a plexus. In all the other cases of this type, the right and left lumbar trunks are distinct. The intestinal trunk joins the caudal extremity of the thoracic duct in 6 instances (figs. 16, 17, 19, 20, 24, 28); the right lumbar lymphatics in 3 instances (figs. 18, 25, 29); the left lumbar lymphatics in 4 instances (figs. 22, 28, 26, 27); and in both the right and left lumbar lymphatics in 1 instance (hie. 21. The caudal extremity of the thoracic duct in 9 instances (figs. 16, 17, 18, 20, 22, 24, 26, 28, 29) presents an ampulliform dilatation, the cisterna chyli. This is absent in 5 instances, and in its place in 4 instances (figs. 21, 23, 25, 27) there is a lymphatic plexus. In 1 instance (fig. 19), there is neither a cisterna chyli nor lymphatic plexus. In addition to the right and left lumber and intestinal branches, the thoracic duct may receive the following branches: (1) col- laterals which drain the intercostal spaces (present in all cases) ; (2) efferent vessels which drain the posterior mediastinal lymph nodes; (3) the left internal jugular trunk in all cases; and (4) the left subclavian trunk in eases 14,15 and 16. The collaterals mentioned above drain, as a rule, more than one intercostal space. There is not, however, a collateral for each intercostal space. The trunks draining the posterior mediastinal nodes had been destroyed in most of the cases. Lymph glands are associated with the thoracic duct in 11 instances (figs. 16, 19, 20, 21, 23, 24, 25, 26, 27, 28, 29). In 9 instances (figs. 17, 18, 19, 21, 24, 25, 26, 28, 29) the duct divides into two branches which unite again after a short distance to form a single trunk. These were called ‘insulae’ by Haller (75). In case 20 (fig. 29) there is a bifurcation of the terminal portion of the duct and each branch presents an ampulliform dilatation, similar to a cisterna chyli. In the abdominal portion, the thoracic ducts of this group lie ventral to the bodies of the first 2 lumbar and 12th thoracic vertebrae and between the crura of the diaphragm or under cover of the right crus. Ventrad they are in relation with the right side of the abdominal aorta. ‘THORACIC DUCT IN MAN 227 ie ' Internal nterna : Jugular Trunk Jugular Trunk Subclavian Trunk Left Internal Jugular Vein eae seinstiscueie Ist va > UL ZSubclavian = i a -Insula Thoracic!” Se -— _ Trunk |st Thoracic Vertebra GE NST ‘a Vertebra=— Ny A Deft Left 4 = Jp \ Subelavian Subclavian— == ii \, = Vein Vein Saat « ‘Left “eau a») 5 Innominate Pa (EE eee Vein a in Collaterals Thoracic BN = Sd (ee: oy Poo ) Dit a a i ii 3= =e SN h Hy) MyM 3 Cisterna VAS Chyli AYN — Righ cy, NY ailaft Lymph Glands Yi 1 Yip Left ~\

Du Thoracic Y= Left : Duct Innominatess Vein Kr ) Collaterals a me ah We ue lies i \ yl: ~My Collaterals SY) ay Fig. 18 Type 6; thoracic duct in a female white subject, age 39. Note the insula at the terminal portion of the duct. Fig. 19 Type 6; thoracic duct in a female white subject, age 50. Note the caudal portion of the duct; there is no cisterna chyli or lymphatic plexus. 228 Internal Left Internal Jugular Vein om LING Ist \ =a (x ? ‘NN ya E—~ Thoracic y= . oy ip pel Vertebra — ASS Left > KS Subclavian —Ah7~—__ \ Vein —© \Left Thoracic Duct Lymph Glands Cisterna : Chyli S| == i nl — UIA \\ ea 7 Intestinal Trunk (= Right PAM. 16 NON Left Lumbar Trunk¥(- >) Lumbar Trunk iM iN ») )) area " Dery ep ht ‘ Uhh Mihi 2 Vi a] 20 terminal portion. Jugular Trunk = eli Subelavian Trunk Left Internal Jugular Vein— 1st Thoracic Vertebra — ch er) Internal Jugular Trunk Subclavian Trunk CRRA i Subclavian > ve oh Vein RO | F : Ir=5 Vein fos Va “Collaterals Thoracic — Duct: = = (Nine } Ee OM z ollaterals a Lymph Gland : ntestinal Trunk : \ 1 Ny os ~ Right: , i ul Left 4) wih Lumbar Trunk Collateral TTY -—> Collaterals 3 is Trunk Left Fig. 22 Type 6; thoracic duct in a female white subject, age 42. Note the abrupt manner in which the right duct crosses over to the left side. Fig. 23 Type 6; thoracic duct in a male white subject, age 43. Note the plexiform arrangement of the caudal portion of the duct and its termination into the posterior wall of the innominate vein. ic = am 7 taternal Internal Left Internal Jugular Vein _Augular Trunk Jugular Trunk. Subclavian 2 ANNA ___ Left Internal Jugular Vein Ist ARN —~ Subclavian ~~ ¢ Z < G (yi Ry \ “Trunk Ist Thoracic Verteprase), Bey ( = UN ere ese S Subclavian aie = ee ‘ ubclavian Vein ee ie Left _~.M c Innominates —“2 Thoracic | = Innominate Vein S = Duct “"v-» Vein Ce 1 Se pas pe eiNe Wie Thoracic —3— Duct ’ a : "i NE Collaterals Lymph Glands \. = Collateral Cs Cisterna Chyli Intestinal > —Lymph Trunk \/ Gland Lumbar 7s Lumbar Trunk (€ Trunk ‘ iN Ri TH Left Why ght > e Lumbar , Je) 1 Lumar Lymphatics Wii a8 fh Lymphatics Mitineee es fe 25 Fig. 24 Type 6; thoracic duct in an adult male white subject. Note the lymph glands along the thoracic portion of the duct and position of the cisterna chyli. Fig. 25 Type 6; thoracic duct in a female white subject, age 51. Note the plexiform arrangement of the caudal portion of the duct and the insula associated with its thoracic portion. 231 252 HENRY K. DAVIS Pea Internal Left Internal Jugular Vein Sei Tuek Left Internal Jugular Vein 4U9H!@r Trunk \ , J YS Subelavian Subclavian Trunk Left Se MI oN Ist Subclavian23— : =" (hen ates Thoracics< Vein 4 = —_ Vertebra “= , : “ — DY (a N Left gS A\ hs > ; 1st eb vila re Se iftisy_b Thoracic en NEF I h(_2 Verteb — So )) . vertebra @ NC | pc Ge: Innominate - a cae V2 3 . : a Thoracic Duct Thoracic a, (WSS wa me—= ‘Collaterals Collaterals Collateral <—Vk a = ise J) Y Me Intestinal Intestinal Trunk Fg i = Lymph ae ff i \ JE ae) Glands SSP A) Cisterna Chyliai™ Sj yy yy Right oI Tai Right icon /) Lumbar WA as aebee Lymphatics YG f fl to lY femipheties eA 26 Fig. 26 Type 6; thor cisterna chyli and the termination of the du Fig. 27 Type 6; thoracic duct in a ma lymph glands associated with the caudal portion of the duct an into the left subclavian vein. acie duct in a male white subject, age 21. Note the et into the left internal jugular vein. le white subject, age 70. Note the d its termination THORACIC DUCT IN MAN Dae Left Internal Jugular Vein Subclavian Internal Jugular Trunk \ 4 Subclavian =A hy) eee 2 JZ Munk Left Keay Ss aN) Trunk ee he ai Subclavian — Sf NY —— 4 -_ aS z~Left Vein oy | == Ze Vv aN meme Cok ubclavian Left Sh Ss PO Gaga Bx. /f i Vein | nnominate. S) Thoracic : ‘mera pe Vein ——— | Vertebra Thoracic < | = pS \\ f ; Peat: Sams ae: ‘ ie a Png (1) >| | SG liqmuy At \ Q, 5 2). IN , age 60. Note the Fig. 28 Type 6; thoracic duct in a male white subject cisterna chyli and the two insulae. Fig. 29 Type 6; thoracic duct in a male white subject cisterna chyli and the two ampulliform dilations at the termin , age 61. Note the al bend of the duct. 234 HENRY K. DAVIS In their thoracic portion, they lie at first within the posterior mediastinum, but cephalad they enter the superior mediastinum. «In the posterior mediastinum, they lie ventrad to the bodies of the 11th to the 5th thoracic vertebrae, and have ventral to them the pericardium, the oesophagus, and the arch of the aorta. The thoracic aorta lies to the left and to the right are the right pleura and the greater azygos vein. The caudal right intercostal arteries pass between them and the bodies of the vertebrae, as does also the terminal portion of the hemiazygos vein. In the superior mediastinum, they rest upon the caudal part of the longus colli muscle, being separated by it from the bodies of the three cephalic thoracic vertebrae. Ventrad they are in relation with the origin of the left subclavian artery and with the vertebral vein; to the left is the pleura and to the right are the oesophagus and the left recurrent laryngeal nerve. The arch of these is in relation caudad with the apex of the left lung and with the left subclavian artery. Dorsad and to the left is the vertebral vein and to the right and ventrad are the left common carotid artery, the left internal jugular vein, and the left vagus nerve. This type of duct belongs clearly to Type 6 of my classification. It occurred in 14 cases out of 22, or in 63.63 per cent. This type of duct is described as normal by all anatomists. Group VII The thoracic duct of this type begins in the abdominal cavity as a single trunk and passes cephalad into the thorax lying on the left side of the aorta. In the thorax it crosses over to the right side and opens into the right angulus venosus. I was unable to find any ducts of this type described in the literature, nor did I find any among my own cases. It seems strange that ducts of this type and also of Type 5 have not been reported, inasmuch as ducts of all the other types have been found. THORACIC DUCT IN MAN Day Group VIII In case 21 (fig. 30) the thoracic duct begins in the abdominal cavity from a plexus of lymphatic vessels and passes cephalad into the thorax. In its course through the thorax, it lies to the right of the aorta, placed between it and the vena azygos major. At the level of the lower third of the 1st thoracic vertebra, the duct divides into two branches which do not become reunited before emptying into the venous system of the right side. The cephalic branch ascends to the 6th cervical vertebra opposite the body of which it begins to incline to the right and divides into two branches which become united again after a course of 20 mm. to form a single trunk which opens into the right internal jugular trunk which opens into the right internal jugular vein a short distance cephalad of its junction with the right subclavian vein. Another branch is given off from this cephalic branch just after it bifur- cates and which opens into the right internal jugular vein cephalad to the opening of the branch just described. The more caudal branch of the thoracic duct passes cephalo-laterad, then iatero- caudad and ventrad to open into the posterior aspect of the right internal jugular vein. The thoracic duct receives the lumbar and intestinal lymphatics in its abdominal portion, collaterals in its abdominal portion, collaterals draining the intercostal spaces in its thoracic portion, and the right internal jugular lymphatic trunk in its cervical portion. There are no vascular peculiarities associated with the duct in this case and there is a left lymphatic duct comparable to the usual right lymphatic duct. Watson (’72), Todd (39), Haller (’75), Cruickshank (’90), and Fleischmann (715) describe cases similar to this, in which the thoracic duct runs its entire course on the right side and opens into the venous system at the base of the neck on the right side. This type of thoracic duct belongs clearly to Type 8 of my classification. It occurred in 1 case out of 22 in my series, or in 4.545 per cent. 236 HENRY K. DAVIS Group IX In case 22 (fig. 31) the thoracic duct begins at the upper level of the 9th thoracic vertebra by the confluence of a plexus of lymphatic vessels. It lies on the left side of the aorta and in this position continues its direction cephalad and opposite the body of the 1st thoracic vertebra begins to incline to the left and then caudad dividing into two branches which open into the left subclavian vein. The caudal branch terminates singly but the cephalic branch divides into three branches just before its termi- nation and opens into the left subclavian vein by three branches. This duct in its abdominal portion receives the lumbar and intestinal lymphatics, in its thoracic portion the collaterals which drain the intercostal spaces and in its cervical portion the left internal jugular and subclavian trunks. Cameron (’02) describes a similar case in which the thoracic duct runs its entire course on the left side of the aorta. This type of duct belongs clearly to Type 9 of my classification. It occurred in 1 case out of 22 in my series, or in 4.545 per cent. VARIATIONS - In 11 cases out of 22, or in 50 per cent, there is a cisterna chy _ present (figs. 14-18, 20, 22, 24, 26, 28, 29). In one of these cases (fig. 14) there is a double cisterna chyli. There is also a double cisterna chyli in case 23 (fig. 832). This case differs from the previous one in that a cisterna chyli is placed on each lumbar trunk and the two lumbar trunks unite to form a single thoracic duct. This case has not been considered among my series be- cause nearly the entire thoracic portion of the duct had been destroyed at the post mortem. Jossifow (’06) reports a similar case in which there was a cisterna chyli on each of the lumbar trunks. Instead of the caudal portion of the duct being dilated as a cisterna chyli it may be represented by a plexus of lymph channels. This condition was found in 10 cases out of the 22, or in 45:45 per cent (figs. 10-13, 21; 23, ‘25, 27; 30; a1).. Ing case out of the 22, or in 4.545 per cent, the thoracic duct is formed by the confluence of the right and left lumbar and intestinal THORACIC DUCT IN MAN Dai Internal Jug la r Trunk \h.. Right Internal Jugular Vein > ; Ee ie GreG TS? mom nema Tall = So Subclavian Jugular Trunk Subclavian ZT pe Vein Left Internal Jugular Vein vA aN | eupelevien Se \ a= aN =F A cama ah HSS ight Ist Thoracic Vertebra —N\) \-’-)) “9 ii o> “Innominate = 4 Po ~ Vein Left =z. HH) = y “hi Y ollaterals . big 3 ” pi Pe Wea LAN Wwe, —— | Lymphatics (are 3 f} 30 Fig. 30 Type 8; thoracic duct in a male white subject, age 55. Note the plexiform arrangement of the caudal portion of the duct and its termination into the right internal jugular vein by three branches. Fig. 31 Type 9; thoracic duct in a male white subject, age 71. Note the position of the duct on the left side, and its quadruple termination into the left subclavian vein. 238 HENRY K. DAVIS lymphatics without the formation of a cisterna chyli and there is no lymphatic plexus (fig. 19). The cisterna chyli is placed opposite the body of the 11th thoracic vertebra in 2 cases out of 22, or in 9.09 per cent (figs. 16 and 22); opposite the bodies of the 11th and 12th thoracic verte- brae in 2 cases out of the 22, or in 9.09 per cent (figs. 15 and 18); opposite the bodies of the 12th thoracic and 1st lumbar vertebrae in 4 cases out of 22, or 18.18 per cent (figs. 14, 17, 20, 29) ; opposite the bodies of the 1st and 2nd lumbar vertebrae in 3 cases out of eal \) Cisterna 2 =a —Chyli Right De IK ©) TY Lumbar Trunk oe ee NF | /} Het »' | Ay Mis tah BP ea / Fig. 32 Thoracic duct in an adult male white subject. Note the cisterna chyli associated with each lumbar trunk. the 22, or in 13.635 per cent (figs. 24, 26, 28); and opposite the bodies of the 2nd and 8rd lumbar vertebrae in 1 case out of 22 or in 4.545 per cent (fig. 14). The intestinal trunk empties into the left lumbar trunk in 7 cases out of 22, or in 31.815 per cent (figs. 11, 12, 16, 22, 23, 26, 29); into the right lumbar trunk in 5 cases out of 22, or in 22.725 per cent (figs. 18, 15, 18, 25, 29); into both the right and left lumbar trunks in 1 case out of 22, or in 4.545 per cent (fig. 21); into the cisterna chyli in 5 cases out of the 22, or in 22.725 per cent (figs. 14, 17, 20, 24, 28); into a lymph plexus in 2 cases out THORACIC DUCT IN MAN 239 of 22, or in 9.09 per cent (figs. 30, 31); and into the thoracic duct in 2 cases out of the 22, or in 9.09 per cent (figs. 10, 19). The point at which the thoracic duct lying on the right side of the aorta begins to incline to the left is subject to some variation. The inclination begins opposite the body of the 3rd thoracic vertebra in 1 case out of 22, or in 4.545 per cent (fig. 22); opposite the body of the 4th thoracic vertebra in 5 cases out of 22, or in 22.725 per cent (figs. 14, 18, 19, 24, 25); opposite the body of the 5th thoracic vertebra in 12 cases out of 22, or in 54.54 per cent (figs. 10-13, 16, 17, 20, 21, 23, 26, 27, 29); and opposite the body of the 6th thoracic vertebra in 2 cases out of 22, or in 9.09 per cent (figs. 15, 18). In 1 case out of 22, or in 4.545 per cent the duct lying on the right side of the aorta did not cross over to the left side (fig. 30) and in 1 case out of 22, or in 4.545 per cent there was no duct on the right side of the aorta (fig. 31). There is a divisiqn of the thoracic duct into two branches which unite again to form a single trunk. This has been termed an ‘insula’ by Haller. One or more insulae occurred in 13 cases out of the 22, or in 59.085 per cent (figs. 10, 12, 18, 15, 17, 18, 20, 2120, 26; 28, 29,30). In 10 cases out of the 22, or in 45.45 per cent (figs. 10, 12, 13, 14,19, 20, 23, 24, 28, 29) lymph glandsaresituated along the thoracic portion of the thoracic duct. According to Sabin (12) lymph glands develop from a lymphatic plexus and as Pensa (’08-’09) remarks, lymph glands may occur anywhere along the course of the thoracic duct. This one may expect, I think, if he recall the early embryonic plexiform arrangement of the thoracic duct. I have observed among my cases the forms of terminations shown in table 1. Terminations of the thoracic duct similar to those described in my cases have been reported by Parsons and Sargent (’09), Wendel (’98), and Verneuil (’53), who cites Boullard’s cases (tables 2 and 3). Table 4 is a comparison of the percentages of the different modes of termination of the thoracic duct as given in tables 1 to 3. From table 4, it will be seen that my results agree quite closely with those of Boullard and with those of Parsons and Sargent. 240 HENRY K. DAVIS SUMMARY The thoracic duct may be double throughout its entire extent, one channel lying on each side of the aorta and opening into the venous system of the corresponding side. This type of duct is similar to the diagram, figure 1. Ducts of this type have been described by four authors. The thoracic duct may be partially doubled and open into the venous system of the left side (figs. 10-15). This type of duct TABLE 1 Termination MODE WHERE CASES | PER CENT FIGURES Singlets A ene | Left angulus venosus 5 | 22.725 | 12, 18, 21, 24, 25 Singles ee eee | Left subclavian 10 | 45.450 | 10, 13-16, 19, | , 20, 22, 27, 28 Single...........| Left internal jugular } 1 | 4.845 26 Sumele...tke ele | Left innominate i! 4.545 23 Woublen cain. | Left subclavian 1 | 4.545 Li Doublepee2e. a2: | 1 branch in I. int. jug. | . | 1 branch in ang. ven. 1), 445 ee 29 riple: peer cse | 2 branches in 1. int. jug. | 1 branch in 1. vertebral 1 | 4.545 11 ripple nee ewes « R. internal jug. 12 Ao545 30° Quadruple.......) Left subclavian 1) 49545 31 TABLE 2 Parsons and Sargent’s cases Termination MODE WHERE | NO. OF CASES | PER CENT Single. foi). 244. | Left internal jugular | 28 | 70.00 Ns 80/83 |< aero oe | Left angulus venosus | 3 | 7.50 Double..........| Left internal jugular | 4 | 10.25 Double..........| Left internal jugular and some | other vein | 2 5.00 Double sae: cece i-br: in 1: int. jug: | 1 br. in 1. subclavian 1 2.50 Quadruple.......; Left int. jug. 1 2.50 Quadruple.......) 1 br. in left int. jug. | 3 br. in left subclavian il 2.50 THORACIC DUCT IN MAN 241 occurred in 6 instances in my series of 22, or in 27.27 per cent. This type of thoracic duct has been described by two authors. The thoracic duct may be partially doubled and open into the venous system of the right side. This type of duct is similar to the diagram, figure 3. One author has described a duct of this type. The thoracic duct may pass cephalad into the thorax on the right side of the aorta as a single duct and divide into two branches, one branch connecting with the venous system of the left side and the other branch with the venous system of the right side. This type of duct is similar to the duct in the diagram, figure 4. Thoracic ducts of this kind have been described by six authors. TABLE 3 Wendel’s cases Termination MODE NO. OF CASES PER CENT SUNG, 3 gd Gab Seed eleetree ERNE Ret a Gace ee aT ec Crt cutee 9 52.941 IDOI DS ie: a coed Moo oc ORS Nees ENC OTTER ee 3 17.647 ‘Wie ONG 5 a5 3516.5 6 eo ss oR Ue ODO clo re en Ren ed Gone 1 5.882 IIo epee ere, race erat ey east 0. '.0.4 02 ana 4 23 .528 Boullard’s cases, reported by Verneuil Termination SHUG 5 cod 5 her etour Stee a orale tes Peer ach aie ae ROMP raee ae eR er 18 74.98 ID Yael aless, sis: oe eR NC ec etc Canela a et ee ee Rn Leer 3 12.49 lini CMP ccm rai Ne Coney crepe cs ala Baia eee: 2 8.33 4.16 Sib Ol Cage rete are ree a Re me da A! ae) Rene ana 1 TABLE 4 PARSONS MODE OF TERMINATION ANDSARGENT’S BOULESED, 5 WEND EE = DAVIS’ CASES: SEE RIEIC CASES: CASES: CASES: Per cent Per cent Per cent Per cent ‘Slval( ta (Cee el a ee OD 74.98 52.941 77.265 ID OWI oa ee eee Res eye rpe ed ox 17.50 12.49 17.647 9.09 ‘Titel Ae eee oon ee 8.33 5.882 9.09 imap e. «4. ed stecky vive oe 5.00 4.545 MRO CAT PO Gaye, ss wisps! oe cuceeecrs edna 23 .528 SHESIOMG| oes eet ee eee roe 4.16 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 2 242 HENRY K. DAVIS The thoracic duct may be single and pass cephalad into the thorax on the right side of the aorta and at about the level of the 5th thoracic vertebra cross over to the left side and open into the venous system of the left side. This type of duct occurred in 14 instances in my series of 22, or in 63.63 per cent (figs. 16-29). This is the most predominant type of thoracic duct and is described as normal by all anatomists. The thoracie duct may lie to the right of the aorta in its entire extent and open into the venous system of the right side. This type of duct occurred in 1 instance in my series of 22, or in 4.545 per cent (fig. 30). Ducts of this type have been described by five authors. The thoracic duct may le to the left of the aorta in its entire extent and open into the venous system of the left side. This type of duct occurred in 1 instance in my series of 22, or in 4.545 per cent (fig. 31). A similar thoracic duct has been de- scribed by one author. Assuming that the thoracic duct is developmentally bilaterally symmetrical, one might expect to find in the adult some cases in which a single duct was situated on the left side of the aorta and divided in the thorax into two branches, one of which would open into the venous system of the left side and the other into the venous system of the right side. This type of duct would be similar to the diagram, figure 5. I found no duct of this type in my own series, nor could I find any described in the literature. Again, assuming that the thoracic duct develops with bilater- al symmetry, it may start in the abdominal cavity as a single duct and pass cephalad into the thorax on the left side of the aorta and at about the level :f the 5th thoracic vertebra cross over to the right side and open into the venous system of the right side. This type of thoracic duct would be similar to the diagram, figure 7. I found no ducts of this type among my own cases nor could I find any described in the literature. It seems strange that no ducts of these last two types have been reported inasmuch as ducts of all the other types have been found and reported. THORACIC DUCT IN MAN 243 A cisterna chyli is present in 50 per cent of my cases (figs. 14-18, 20, 22, 24, 26, 28, 29) from which results it is evident that a cisterna chyli is not present as often as one would suspect from the descriptions of the thoracic duct in modern anatomical textbooks. In 59.085 per cent of my cases, there is an insula associated with the thoracic duct (figs. 10, 12, 13, 15, 17, 18, 20, 21, 25, 26, 28, 29, 30). Haller (75) considered this the normal condition. The thoracic duct terminates singly in 77.265 per cent of my cases (figs. 10-16, 18-28); doubly, in 9.09 per cent of the cases (figs. 17, 29); triply, in 9.09 per cent of the cases (figs. 11, 30; and quadruply, in 4.545 per cent of the cases (fig. 31). The thoracic duct terminated in the left subclavian vein in 59.085 per cent of my cases (figs. 10, 13-17, 19, 20, 22, 27, 28, 31); in the left innominate in 4.545 per cent of the cases (fig. 28); in the left angulus venosus in 22.725 per cent of the cases (figs. 12, 18, 21, 24, 25); in the left internal jugular in 4.545 per cent of the cases (fig. 26); in the right internal jugular in 4.545 per cent of the cases (fig. 30); in the left internal jugular and left angulus venosus in 4.545 per cent of the cases (fig. 29); and in the left internal jugular and left vertebral vein in 4.545 per cent of the cases (fig. 11). In conclusion, it is a pleasure to thank Dr. Abram T. Kerr for many valuable suggestions and continued interest throughout the course of this work. BIBLIOGRAPHY BreESCHET, G. 18386 Le systeme lymphatique. Buruer, C.8. 1903 Anabnormal thoracic duct. Journal of Medical Research, Boston, vol. 10, p. 154. CaMERON, 8S. 1903 Case of right aortic arch with abnormal disposition of the left innominate vein and thoracic duct. Lancet, London, vol. 2, p. 670. Cousin, G. 1898 Anomalies du canal thoracique. Bull. Soc. Anat., Paris, tome 73, p. 334. CRUICKSHANK, W. 1790 Anatomy of the absorbing vessels. DIEMERBROECK, T. 1685 Opera omnia anatomica et medica. EKustacuius, B. 1564 Opuscula anatomica, Venetiis. FLEISCHMANN, G. 1815 Leichen6ffnungen. 244 HENRY K. DAVIS dardint Hauer, A. 1775 Elementa physiologiae corporis humani. Tome 8. 1746 Disputationes anatomicae selectae. Tome 1. Homer, D. 1737 Litterarium ad rei commercium medicae et scientiae. Norim- berg, p. 162. Huntinaton, G. 8. 1911 The anatomy and development of the systemic lym- phatic vessels in the domestic cat. Josstrow, G. M. 1906 Der Anfang des Ductus Thoracicus und dessen Erweite- rung. Archiv fiir Anat. und Physiol., Anat. Abt., Tafl. 1. Lautu, E. A. 1885 Nouveau manuel de l’anatomiste. Ler, A. B. 1905 The microtomists vade-mecum. Lewis, F. T. 1906 The ties of the lymphatic system in rabbits. Am. Jour. Anat., vol. 5, p. 95. Lower, R. 1680 We. ae Nuun,, A. 1849 Untersuchungen und Beobachtungen a. d. der Gebiete der Anatomie. Heidelberg. Parsons, F. C., and Sarcent, P. W.G. 1909 Termination of the thoracic duct. Lancet, London, vol. 1, p. 1178 von PaTRUBAN, C. E. 1844 Medicinische Jahrbiicher des Kais. Koénigl. Oester- reichischen Stattes. Bd. 39, Seite 24. Precqurt, J. 1661 Experimenta nova anatomica. Amsterdam. Pensa, A. 1908-09 Studio sulla morphologia e sulla topographia della eysterna chili e del ductus thoracicus nell’ uomo ed in altri Mammifera. Ricerche fatte nel Laboratoria di Anatomia Normale della R. Universita di Roma, vol. 14, p. 1. Sasin, F. R. 1901-02 On the origin of the lymphatic system from the veins and the development of the lymph hearts and the thoracic duct in the pig. Am. Jour. Anat., vol. 1, p. 367. 1909 The lymphatic system in human embryos with a consideration of the morphology of the system as a whole. Am. Jour. Anat., vol. 9, p. 41. 1912. The development of the lymphatic system. Keibel and Mall’s human embryology, vol. 2, p. 734. Saua, L. 1899-00 Sullo sviluppo dei cuori limfatici e dotti toracici nell’ embrione di pollo. Ricerche fatte nel Laboratoria di Anatomia Normale della R. Universita di Roma. Table 14, fig. 16. Topp, R. B. 18389 Cyclopedia of anatomy and physiology, vol. 3, p. 232. VERNEUIL, A. G. 1853 Le systeme veineux. Paris. Watson, M. 1872 Note on the termination of the thoracic duct on the right side. Jour. Anat. and Physiol., vol. 6, p. 427. WENDEL, W. 1898 Ueber die Verletzung des Ductus Thoracicus am Halse und ihre Heilungs-méglichkeit. Deutsche Zeitschrift fiir Chirurgie, Leip- zig. Bd. 48, p. 4387. Winstow, M. 1866 Exposition anatomique de la structure du corps humain, tome 3. THE HISTOGENESIS OF THE SELACHIAN LIVER RICHARD E. SCAMMON Institute of Anatomy, University of Minnesota FORTY-FIVE FIGURES (SEVEN PLATES) CONTENTS ee OULU GION eer eta a Ks Sei atl eee Biro Re rcrah eh Hd og aces o'n! Sha taelehine a old 245 Wevelapment or the hepatic Gylimgens.: .0\-q. gees: a2. esa es ets k cede wn dete 246 ley LUTE REE Toe aa a Oe Boe Sire a PO coat hale At J Secs a a a 246 Pe Mablyed eve lOpme»mites cnt: luce ty stata ae eRe SNeE aks foie ahi cases) Gat eyarepeies 249 Som NCRPLOCESSUOL ANASTOMOSIS Hs. 4a eae nen eet cere = 8 }spatcs (a sehen e ceepauetere 253 dhs LDP ysTeroats FENIeMG) 0} 60% = 0) Ney ve ts Pe ko em SUEUR SS 2. 5c cr ne 258 Ae Grow bh Ofeiheyhepaticen et woke rerio 26 ces ss nee ce epee 258 b. Changes in size and structure of the hepatic cylinders........ 261 5. Development of the hepatic tubules in different forms of selachians.. 266 Development of the minor rami of the hepatic ducts........................ 269 evelopment of the hepatic mesenchyma. .../o20 82 28 .. ki kes eee sees ee Bae 27 MistenveOrghe mepatic SIMUSHIOS. ho... s'.\\ cea. as a hoa knee «Ante nue 282 Summary... “fe ee AU TS a0 SOU rnc 5 Se ct AP 294 [SYS INSpa ea yo) SRE eR 8 lh on ie Ro) OT ae Ae oe 299 INTRODUCTION The structure of the adult selachian liver is far removed from that type which is generally considered characteristic of the organ in ertebrates. This variance from the common verte- brate type is seen in the great accumulation of fat in the hepatic cells, in the comparatively slight development of the bile duct system and in the absence of lobulation of the kind generally found in higher vertebrates... However, these pecularities which 1 The histology of the adult elasmobranch liver was first briefly described by Leydig (’51) from observations on Chimaera. He published a more complete account dealing with several forms of selachians in 1853. Later descriptions of the general histology of the adult liver have been given by Shore and Jones (89), Pilliet (90), and Holm (’97). Deflandre (’05) has investigated the fat content of the hepatic cells, and Monti (’98) has studied the bile capillaries by the Golgi method. 245 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, NO. 3 MARCH, 1915 246 RICHARD E. SCAMMON distinguish the selachian liver are not manifested until a com- paratively late stage in the development of the organ. In earlier stages characters common to the liver in all vertebrates, but which are often masked or modified in higher forms, are shown with unusual clearness. It is chiefly with these more fundamental characters such as the formation and anastomoses of the hepatic cylinders, the differentiation of the minor bile | ducts and the relation of the parenchymatous and vascular structures in the liver, that this paper has to do. The specific characters of the selachian liver, which have been mentioned, have been considered only incidentally. The main material employed in this study consisted of embryos of Squalus acanthias, but specimens of Raia batis, Torpedo ocellata, Mustelus canis, Mustelus laevis and Squatina angelus have been used for supplemental and comparative work. For a large part of the Acanthias material, and for the specimens of Mustelus laevis, I am particularly indebted to the late Dr. Charles S. Minot, who not only permitted the removal of numerous series from the Harvard Embryological Collection, but also provided special material for use in this work. DEVELOPMENT OF THE HEPATIC CYLINDERS 1. Literature Our conception of the glandular structure of the vertebrate liver rests upon a large number of observations made mainly in the first half of the last century and culminating in the work of Eberth (66) and Herring (’72). Since that time it has been recognized that the liver is a compound gland with a more or less regularly branching system of ducts and with a terminal network of anastomosing end pieces, and our knowledge of the details of this network has been greatly extended by the Golgi method in the hands of Retzius (’92), Hendrickson (’96) and others. It is generally stated that the anastomotic type of liver is a modification of the compound branching gland type: Of this the best proof is the phylogenetic one, for in the lower cyclos- ee ee * ne ye HISTOGENESIS OF THE LIVER 247 tomes, the liver is indeed a true compound branching gland, as has been shown by the work of Retzius (92), Holm (97), Cole (’13) and others. In Petromyzon, however, anastomoses occur occasionally in the embryo (Brachet ’97) and are very numerous in the adult (Renaut 797). On the side of embryology little evidence of this modification has been offered, although the statement that the liver arises in many vertebrate embryos as a branching gland, and that it takes on its adult reticular structure through the anastomoses of its end pieces, is common enough in texts. But an examination of the literature shows no more complete account of this early period of the histogenesis of the liver than that given by Remak (’55) excepting Hilton’s work (’03) on the pig which has hardly received the attention that it deserves. The study of the actual course of the anasto- mosis of the end pieces has also been neglected, and the only clear statement to be found regarding this process is not in connection with the liver at all but in Laguesse’s (’96) description of the histogenesis of the pancreas where he found numerous anastomoses at different stages. The literature of the histogenesis of the liver in selachians is scanty and has been to a large extent reviewed in the com- prehensive works of Oppel (’00) and Fiessinger (711). Balfour’s (’76) account was the first to appear and as it has formed the basis of all later descriptions, it is given here in full: By stage K the hepatic diverticula have begun to bud out a number of small hollow knobs. These rapidly increase in length and number and form the so-called hepatic cylinders. They anastomose and unite -together so that by stage L there is constructed a regular network. As the cylinders increase in length their lumen becomes very small] but appears never to vanish. Hammar (’93) in the course of a study devoted particularly to the development of the larger hepatic structures illustrates the development of the hepatic cylinders in two models of the liver of Torpedo embryos corresponding in development to Balfour’s stages L and M. He makes no comment upon them aside from stating that the trabeculae increase in number and decrease in caliber in the period intervening between the two stages. 248 RICHARD E. SCAMMON Brachet (95) studying Torpedo embryos determined with accuracy the area of the hepatic diverticulum which gives rise to the hepatic tubules, and confirmed Balfour’s description of the structure of the hepatic cylinders in later stages. Holm (’97) figured and described very briefly sections of the liver of two Scyllium embryos of advanced stages. Most of our later information concerning the embryonic hepatic parenchyma comes from the studies of Braus (’96) on em- bryos of Acanthias, Spinax and Scyllium. Confirming Balfour’s observations in regard to early stages he noted a complete and regular anastomosis of the tubules in Acanthias embryos of 38 mm. Here the tubules were of even size and consisted in cross section of seven cells surrounding a lumen of variable size but distinctly larger than that seen in the adult. The hepatic cells of this stage were free from fat. In older embryos of Spinax and Acanthias the cells were fat laden. Braus saw no side branches nor blind endings of any hepatic capillaries. Choronshitzky (’00), studying Torpedo, found several secondary hollow outgrowths from the hepatic pouch in his ‘Stadium II,” which corresponds approximately to Balfour’s stage K, and to the Normal-plate Nos. 22-24. In ‘Stadium III” which is represented by considerably older embryos, the liver pouches give rise to a number of small hollow buds, the cavities of which communicate with that of the pouch. ‘‘Die Leber macht im alleemeinen den Eindruck einer verzweigten Driise.” By “Sta- dium IV” the hollow buds have been transformed into much branched hepatic trabeculae which contain no traces of lumina. Choronshitzky’s opinion of the mechanical influence of the blood vessels on the formation of the hepatic tubules will be discussed in a later part of this paper. Minot (’00) in discussing the development of sinusoids, men- tions the presence of the first short hepatic cylinders in an Acan- thias embryo of 11.5 mm. and speaks of their anastomosis and growth in older stages. He noted the interesting retardment of development which is to be seen in later stages in the caudal tip of the liver as compared with the cardiac end. > HISTOGENESIS OF THE LIVER 249 Debeyre (’09) made use of observations upon the development of the hepatic cylinders in Acanthias to lay the ghost of the theory of the mesodermal origin of the hepatic parenchyma, which had been again raised a short time before by Géraudel (07). He gives no complete or detailed history of the cylinders but notes, with illustrations, their general appearance in embryos 16, 22, and 30 mm. in length, respectively. In the latter he recognized the beginnning of a period of pronounced increase in the diameter of the cylinders. Debeyre noted the presence of numerous granules in the apices of the hepatic cells and bases upon this the interesting suggestion that the liver may serve as an organ of internal secretion during a part of embryonic life. 2. Early development of the hepatic tubules In this account the structures which have been variously termed hepatic cords, trabeculae, cylinders and tubules will be spoken of as tubules as long as they remain as portions of simple or branching unanastomosed glands. The term hepatic cylinders will be employed for the same structures after the process of anastomosis has taken place. The exact time when the anlagen of the hepatic tubules first make their appearance is somewhat variable. In general they are first to be seen in embryos from 7.5 to 9 mm. in length, being somewhat younger than Balfour’s stage K? and corresponding to numbers 22 and 23 of the Normal plate series. Such embryos have from fifty to sixty-five segments and four or five pairs of gill pouches of which the anterior three or four may open extern- ally. The spiral valve is in the process of formation, making at this time one or two complete turns of the intestine and the vitelline duct is reduced to a short wide canal. The form of liver anlage at this stage is represented somewhat diagrammati- cally in figure 1. The organ consists of a ventral median pouch from the foregut just anterior to the vitelline duct. The anterior 2 In correlating embryos with stages of Balfour’s series only the general develop- ment of the embryo has been considered and not the state of development of the organ under discussion. 250 RICHARD E. SCAMMON part of this pouch is full and rounded and may be termed the pars hepatica mediana. From the median pouch spring two large lateral pouches which form together the pars hepatica lateralis. ‘In stages just preceding the appearance of the tubule anlagen the lateral pouches are smooth and globose and project outward almost at right angles to the median hepatic pouch. At the time when the tubules are formed, however, the lateral pouches are flattened : transversely and have entered upon a pronounced dorsal growth. Connected with the liver pouch above and in front and with the anterior wall of the yolk stalk behind is a small thick walled sac, the anlage of the gall bladder. The hepatic tubules take Fore gut Fig. 1 Lateral view of a semi-diagrammatic reconstruction of the foregut and liver of an Acanthias embryo 9 mm. long. The areas represented in stipple give rise to hepatic tubules. origin entirely from the pars hepatica mediana and the pars he- patica lateralis. These areas are indicated in stipple in figure 1. The tubule anlagen arise in two forms: as slight longitudinal ridges upon the lower part of the outer surface of the lateral pouches and as very small irregularities of the dorsal margins of the same structures. When first observed the longitudinal ridges are two to four in number on either side. They extend almost the entire length of the lateral pouches and are distinctly separated by shallow lateral furrows. Sometimes these ridges may be subdivided longitudinally at their ends. Figure 17 is a view from the left side and below of the liver of an embryo 7.5 HISTOGENESIS OF THE LIVER 251 mm. in length (S.C. 14). This specimen bears two ridges on the left hepatic pouch, but the right pouch is entirely smooth. The irregularities of the margin of the hepatic pouches, which also form tubules, are at first so ill defined as to be scarcely noticeable unless reconstructed. Then it is seen that the formerly straight dorsal margin has a wavy contour. The tubule ridges rapidly increase both in size and number. New ridges appear ventral and mesial to the earlier ones and extend to the base of the lateral pouches and upon the ventral surface of the pars hepatica medialis of the median liver pouch. No new ridges appear dorsal to the first ones and I thin that no new ridges arise between older ones. By the time the embryo reaches a length of 10 to 12 mm. each lateral pouch wall and lateral half of the pars hepatica medialis bears seven to ten tubule ridges. This is the total number formed on either side, and when tubules thereafter arise from the pouch directly, they do so as individual tubules and not in the form of tubule ridges. ’ Figure 18 is a view from the left side and below of a reconstruction of the liver of an Acanthias embryo 10 mm. in length (S.C. 20) showing the later form of the tubule ridges and the beginning of the differentiation of tubules from them. Before all the tubule ridges are formed the ones which first develop are broken up by transverse or oblique furrows into rather irregular rows of low mound like elevations. These elevations are the anlagen of the individual hepatic tubules and may be seen in figure 18 referred to above. They are semi- circular or nearly so in cross section and nearly twice as long as broad, their greater length being always directed antero-poste- riorly. Almost immediately the tubules begin an active outward growth and each is differentiated into a distal extremity which often is large and pouch like, and a’ more slender proximal stalk which is connected with the hepatic pouch and which is circular in cross section. The further growth of the tubules takes place by the formation of tubules of the second order from the distal 3 In designating embryos the following abbreviations will be employed: H.E.C., Harvard Embryological Collection; K.U.E.C., Embryological Collection, De- partment of Zoology, University of Kansas; S.C., author’s collection. 252 RICHARD E. SCAMMON expansions of the primary ones. The secondary tubules are short conical projections rarely over one and one-half times as long as their greatest diameter. They arise from the sides of the distal expansion and almost always extend out at right angles to the axis of the primary tubule. Those primary tubules which arise from the sides of the hepatic pouches and are packed in among their fellows commonly assume a T-shaped form with the limbs of the T directed antero-posteriorly. ‘The primary tubules which arise from the margins of the hepatic pouches become somewhat larger than those formed from the tubule ridges and give rise to from three to seven secondary tubules from their expanded distal chambers. These secondary branches may again sub- divide into branches of the third order and upon these in turn there may occasionally be found nipple-like projections which represent the fourth order of tubules. In the large majority of cases, however, anastomosis takes place before tubules of the fourth or even the third order are formed. Figures 19 and 20 are of wax reconstructions of tubules from the lateral hepatic — pouch of an embryo 13.3 mm. in length (8.°C. No. 18). In figure 19 one sees the beginning of the outpouching of the distal ex- pansion into secondary tubules. These structures are well marked projections in the older tubule shown in figure 20. Figures 22 and 23 are two views of a hepatic tubule from the dorsal margin of the left hepatic pouch of an embryo 15 mm. long (H.E.C. 227). The latter figure shows tubules of the fourth order. Figure 21 is a wax reconstruction of two young primary tubules from the ventral surface of the pars hepatica medialis of the same 15 mm. embryo. One of these shows the beginning of tubules of the second order. ~ When the tubule ridges first appear on the hepatic pouches they are due to the increased thickness of the epithelium in these places and not to an evagination of the pouch wall. In cross sections of the ridges (see fig. 28, a cross section of the lateral wall of the left hepatic pouch of an embryo of 8 mm. K.U.E.C. 542) it is noticeable that the nuclei which in other parts of the wall are arranged in two interlocking rows tend to be reduced HISTOGENESIS OF THE LIVER 253 to a single row which lies near the external surface of the epithe- lium. ‘There seems to be but a single layer of very high columnar cells at this point but the cell boundaries are not very clear. There is always a broad clear zone of cytoplasm towards the lumen of the pouch opposite the tubule ridge. Mitotic figures do not occur in the ridges but are frequent in the zones of epithe- lium between them. Following the formation of the ridge there appears a very shallow trench on the internal surface of the epithelium. The lumina of the individual tubules appear as the tubules themselves are differentiated through the breaking up of the ridges into rounded anlagen. One then finds in each tubule anlage a narrow slit-like cleft passing between the cell walls at right angles to the pouch cavity. This is shown in figure 29. Thus at first the tubule cells do not lie at right angles to the tubule lumen but nearly parallel toit. Later the cells assume an oblique position (see again fig. 29) and finally come to lie in the typical radiating position in relation to the lumen (fig. 30). As the tubules develop the cells become shorter and their nuclei change from elongately ovoid to nearly spherical bodies. 3. The anastomosis of the hepatic tubules The liver is transformed from a gland of the branching to one of the reticular type by the anastomosis of its end pieces. The process begins in Acanthias in embryos from 12 to 15 mm. in length. Such embryos have from 70 to 85 pairs of somites and correspond roughly with Balfour’s stages L and K and with Nos. 25-26 of the Normal plate series. The tubules arising from the dorsal portions of the lateral liver pouches precede in their differentiation those of other regions and it is generally among these dorsally placed tubules that anastomoses are first found. Later the tubules of the lower parts of the lateral pouches and finally those of the pars hepatica medialis enter upon the process. Variants from this general plan of procedure are not uncommon. Figure 2, of a sagittal section of the left hepatic pouch of an embryo 14 mm. long (S.C. 30) shows how general anastomoses are when once they begin in a given region. 254. RICHARD E. SCAMMON No particular degree of differentiation seems necessary before a*tubule takes part in an anastomosis, and when this process begins in any region both branched and simple tubules fuse indifferently. The commonest form of anastomosis is that established by the end to end fusion of tubules, but tubules may Fig. 2 Sagittal section of the anastomosing tubules of the left hepatic pouch of an Acanthias embryo 14 mm. long (8.C. 30.) X 100. also join end to side or side to side. The last type is much com- moner in later stages when more opportunities are offered for this form of contact. Anastomoses also seem to take place with equal frequency between tubules formerly quite separate or between minor branches of the same tubule when it is possible for such branches to come in contact. HISTOGENESIS OF THE LIVER 255 Figure 24 is of a wax reconstruction of three anastomosing tubules from the left hepatic pouch of an embryo 15 mm. in length (H.E.C. 227). The pouch wall from which the tubules spring is cut away squarely around the base of each tubule. Tubules A and B are completely anastomosed end to end but there is still a distinct constriction at the plane where they have joined and the lumen within is considerably narrower than in the bodies of the tubules. Although not clearly shown in the figure both A and B are T-shaped and in each case the other tip of the tubule ends freely without anastomosis. Tubule C is a compound one and the subdivision which passes over to join with B is a branch of the third order. There is as yet no actual anastomosis between C’ and B, but the epithelial walls of the two are in direct contact. The lack of any account of the histology of anastomosis seems to warrant a rather full description of the process here. The process may be rather arbitrarily divided into three steps or stages. In the first step the tubules come in actual contact; in the second there occurs the fusion of their walls and a re- arrangement of the cells forming them, and in the third there is an establishment of a connecting channel between the two original lumina. These stages of anastomosis can be observed in the liver of Acanthias embryos of any length from 13 to 45 mm. They are more easily followed, however, in young specimens. The figures which are used here to illustrate the process have been taken from embryos under 20 mm. in length. In each case the sections have been followed through and the tubules involved have been reconstructed to make sure that the picture presented was not due to an oblique plane of section or a misinterpretation of a segment of a single complicated tubule. Early stages of contact of the anastomosing tubules are illus- trated by figure 30 which is a transverse section of the left hepatic - pouch of an embryo 13.3 mm. long (8.C. 18). Numerous tubules extend outward from the lateral wall of the hepatic pouch. Of these the larger number are still in the form of simple tubules which are expanded distally but a few have entered upon anasto- 256 RICHARD E. SCAMMON mosis. The blood vessels pass among the tubules covering them in part with a thin film of endothelium which is interrupted at some places by the attachment of strands of mesenchyma and at others by the apposition of the distal surfaces of tubules with the insheathing layer of splanchnic mesothelium. Whether this endothelial covering forms an absolutely continuous partition without fenestra between the blood and the tubule epithelium cannot be definitely determined without injected specimens, but there seem to be places where the separation is not complete. Anastomosis is inaugurated by the contact of the involved tubules. At first a few mesenchymal or endothelial cells may separate the tubules but these are apparently pressed to either side so that the entodermal cells soon lie in actual contact. Often at first only four or five tubule cells actually meet, but shortly there is formed an area of contact which generally is not quite so large as the caliber of either tubule involved. Sometimes the cells of one tubule indent the wall of the consort but this is not the common rule. There are no basement membranes about the hepatic tubules, but the basal margins of the cells seem a little thickened so that for a time after contact the line between the cells of the two anastomosing tubules is still distinct. With the disappearance of this line the tubules may be considered as fairly fused. The connection between the two fused tubules is often drawn . out a little forming a short stalk or bridge between them. This will be termed here the connecting stalk. An indentation upon the external surface of this stalk indicates the line of fusion of the tubules. The connecting stalk, when present, is generally a little less in diameter than are either of the tubules and the boundaries -of the cells forming it are not clear. One can follow the cells, however, by the position of their nuclei. Although at first sight the nuclei appear scattered without order in the stalk or at the point of anastomosis, a little study enables one to determine with some accuracy which nuclei are contributed by each tubule. The two rows of nuclei approach one another and may interlock but their radial arrangement in regard to the lumina of the two original tubules is not lost at first. They then pass to one side or the other of the connecting stalk, leaving a clear cytoplasmic HISTOGENESIS OF THE LIVER 257 core in which cell boundaries are faintly distinguishable. In so doing the axes of the nuclei rotate through an angle of 90° so that instead of being parallel to the long axis of the connecting stalk as at first, they are now at right angles toit. Such rotation can be determined, of course, only in young tubules where the nuclei are oval and not circular in section. Figure 26, drawn from an embryo 19 mm. in length (S.C. 3), shows an early step in these changes. Figure 3A is a graphic reconstruction of the tubules shown in cross section in figure 26. In this section the rows of nuclei belonging to the two tubules involved are distinguishable although the process of migration of the nuclei to the sides of the connecting piece is clearly under way. In figure 36 of a later stage from an embryo 14 mm. in length (S.C. 30), all the nuclei with the exception of one have passed to the sides of the connecting stalk. In following the course of the nuclei in tubule anastomosis one is but tracing the movements of the cells in which they are contained, for it is hardly to be considered that the nuclei shift their axes within the cells, and moreover the few faint cell bound- aries which may be made out show the same changes in position as do the nuclei. The hepatic cells of the connecting stalk have shifted through an are of about 90° and when a lumen is estab- lished through the center of the connecting stalk it is bounded at least in greater part by the same cell surfaces which were presented to the lumina in the original tubules. In other words, while the tubule cells involved in anastomosis have shifted in position, their surfaces and their axes will bear the same relation to the new lumen which they did to the former one. Their long axes will be at right angles to the lumen while the inner and outer surfaces of the cell remain constant in both the original position in the simple tubule and the later position in the anastomotic segment. The polarity of the cell, in the sense of the term as used by Rabl (’88, ’90), is not disturbed by anastomosis. The lumen of the anastomosis is formed by clefts which ex- tend out from the lumina of the formerly simple tubules. These clefts are at first small and irregular. They pass between the rather irregular borders of the radially arranged cells of the con- 258 RICHARD E. SCAMMON necting stalk or plate, and, meeting, become confluent. It is not until some time after these clefts have joined that the lumen of the anastomosis acquires its full size and regularity. A late stage in the establishment of this connecting channel is shown in figure 27 from an embryo 19 mm. long (8.C. 3). Here the clefts are stil! separated by a single cell. Figure 3 B is a graphic reconstruction of the tubules involved in this anastomosis. Fig. 3 Graphic reconstructions of anastomosing hepatic tubules in Acanthias. Detailed drawings of the sections indicated with dotted lines will be found in plate 2. A, tubules from an embryo 19 mm. long (S.C. 8) (see fig. 26). B, tubules from an embryo 19 mm. long (S.C. 2) (see fig. 27). > 100. 4. Later history of the hepatic cylinders a. Growth of the hepatic network. Immediately after anasto- mosis has occurred the liver increases in size very rapidly. At first this growth is due almost entirely to the increase in number of the hepatic cylinders, but later the greater portion is brought about by the tremendous enlargement of the hepatic sinusoids. After anastomosis new cylinders are added to the existing net- work in three ways: by the formation of tubules from the remains of the hepatic pouches, by the formation of blind sprouts or buds from the sides of cylinders forming the network and by the production of new cylinders at the periphery of the network from the cylinders located there. These methods of addition to the hepatic network cease in the order given and the proportional amount contributed to the network by the several methods is in inverse order to that in which they are listed above. HISTOGENESIS OF THE LIVER 259 The direct production of new hepatic cylinders from the hepatic pouches contributes but little to the bulk of the hepatic network and continues for only a short period after anastomoses are formed. As described in section 2, there are numerous tubules which arise at the bases of or between tubules which have previously sprung from the hepatic pouches. These younger tubules are at first small nipple shaped elevations arising singly and not from ridges. So long as the earlier tubules remain simple the later tubules follow them closely in their development, but when anastomoses become common among the older tubules and the sinusoidal circulation is well established, the younger tubules anastomose almost immediately after their formation either with one another or with cylinders of the already estab- lished network. In later stages hepatic tubules may sometimes arise as loops, the ends of which are attached to the wail of the hepatic pouch. By the time the embryo reaches a length of approximately 20mm. (Normal plate Nos. 27-28, Balfour’s stage N) the hepatic pouches are transformed into veritable hepatic ducts and thereafter no new hepatic tubules arise from them. Hepatic cylinders, as has been seen, give rise to secondary buds while still in the form of single tubules, and this budding process continues long after anastomosis. The process can be most clearly demonstrated by the means of thick sections of which figure 41 from an embryo 24 mm. in length is an example. 4 The method of preparation of the thick sections used in this study was as follows: The embryo was infiltrated with celloidin and cemented to a piece of infiltrated spleen or liver which in turn was fastened to a fiber block. The speci- men was then cleared by Gilson’s method, which makes the block almost trang- parent. The block was then placed in the microtome clamp and a strong beam of light from a condenser directed upon the object. A Greenough binocular micro- scope equipped with low power lenses was set up over the object. With this arrangement it was possible to follow in detail the process of section cutting. Sections were then carefully cut away until the exact region desired was reached. By focusing with the binocular it was possible to determine the approximate thickness of the section needed to just include the desired structures, and this section was then removed with a single cut. Sections made in this way are often superior to reconstructions for the study of the form of very small structures. They are best stained in a very dilute carmin solution and cleared for a long period in.cedar oil, after which they can be observed with the binocular micro- scope, or better with the aid of an Abby binocular eyepiece. 260 RICHARD E. SCAMMON In the area of approximately 1 mm. represented here there are five blind buds projecting from the cylinders into the sinusoids. This method of addition to the network continues for a con- siderable period. I have found no traces of new buds in the body of the liver after the great increase in size of the cylinders when the embryo reaches a length of about 40 mm. However, in the portions of the liver which are the last to be formed, 1.e., the dor- sal margins and the posterior tips of the lateral lobes, this method of cylinder formation continues until the embryo reaches a length of 50 to 60 mm. While in the earlier stages of the development of the hepatic network the increase comes perhaps equally from peripheral and interstitial growth, in later stages the latter method is by far the more important. The hepatic network terminates periph- erally in a large number of blind knobs which by their growth and division give rise to a large amount of hepatic tissue. The cells of these terminal knobs remain in a comparatively undiffenti- ated condition while those of the more central part of the network are undergoing rapid changes in structure. Figure 43 shows a small portion of the tip of the lateral lobe of an embryo 20 mm. long. The mesothelium covering the liver has been removed. Here the terminal knobs are seen projecting from the general network and are often attached to the mesothelial sheath by strands of mesenchyma. This specimen was prepared by cutting a thick celloidin section of the region desired by the method already described. The celloidin was then dissolved away and the mesothelial covering stripped from the fragment with fine forceps. ge The peripheral addition to the hepatic network takes place at first over the entire surface of the liver. But like the inter- stitial method of addition it is later limited to the tips of the lateral lobes and to the dorsal margins of both the body and lateral lobes of the liver which at a comparatively late stage grow rapidly upward between the stomach and the lateral body walls. In embryos 60 mm. in length these areas are much re- stricted and they cannot be seen in an embryo 80 mm. long. HISTOGENESIS OF THE LIVER 261 In the rapidly growing parts of the liver the cylinders often terminate peripherally in expanded end-bulbs or vesicles which contain a lumen many times the diameter of that of the typical cylinder. Such vesicles are found in embryos from 25 to 40 mm. in length but not in older specimens. Their walls consist of a low columnar or cuboidal epithelium. Figure 4 is a cross section of one of these structures from the lateral lobe of an embryo 36.6 mm. long (S.C. 10). The significance of these vesicles is “unknown to me. Fig. 4 Terminal vesicle of a hepatic cylinder from the lateral lobe of an Acan- thias embryo 36.6 mm. long (8.C. 10). * 400. JL, lumen of hepatic cylinder; V, terminal vesicle; X, side branch from central lumen of cylinder. b. Changes in the size and structure of the hepatic cylinders. When the hepatic tubules enter upon anastomoses they are irregular in form and of variable caliber. The lumina of the tubules are large and irregular and are generally surrounded in cross section by 14 or 15 eells, if one may judge from the number of nuclei present, for cell boundaries are often indistinct at this time. Table 1 shows some of the changes which take place in the ‘course of the later development of the cylinders. The measure- ments and cell counts given in this table are in each case averages determined from twenty fair cross sections of cylinders of the THE AMBRICAN JOURNAL OF ANATOMY, VOL. 17, No.3 262 é RICHARD E. SCAMMON posterior part of the median lobe of the liver. ‘The cross sections of cylinders were taken at random from this region, except that those lying in the rapidly growing peripheral zone were avoided in each case. In the first specimen of the series the number of cells bounding the lumen in cross section averages 12.5. The diameter of the tubules has become more uniform and averages 58 micra. Ex- amples of such tubules are illustrated in figure 31. At about this time the liver begins to increase greatly in size. This growth is — due in part to the actual increase in number of hepatic cells, as TABLE 1 Measurements of the hepatic cylinders in Acanthias embryos LENGTH NP SEEANEH DA ISUITAE) AVERAGE DIAMETER PaeeT? «| promgasignemon’ (| Paauarias 15.0 | 12.5 58.0 20.6 10.0 50.0 25.0 8.0 37.6 28.0 7.0 35.6 33.0 | 6.6 37.0 37.0 | 6.0 38.5 47.0 | 5.2 40.0 50.0 4.6 46.0 60.0 | 5.3 39.0 80.0 | 5.0 43.0 | 3.7 54.0 95.0 is indicated by the presence of numerous mitotic figures in the hepatic cylinders, but a much greater part of the increment is due to the establishment of the huge hepatic sinusoids. With this increase in size of the sinusoids the tubules are distinctly reduced in size, their average diameter dropping from 58 micra in an embryo 15 mm. in length to 37.6 micra in one 25 mm. and 35.6, in one 28 mm. in length. This reduction in diameter may be due somewhat to the decrease in the size of the lumen, but is’ caused mainly by the actual decrease in the number of cells surrounding the lumen at any one plane. The process is a con- tinuous one after anastomosis is established, but is more rapid HISTOGENESIS OF THE LIVER 263 at first, dropping from an average of 12.5 in an embryo of 15 mm. to 7in one of 28mm. I think that this rapid and early reduction is due to the stretching of the tubules caused by the dilatation of the sinusoids among them. The tubules are attached to each other and to the mesothelial wall of the liver by strands of mesen- chyma which may aid in the process by pulling upon the cylinders as the mesothelial wall is rapidly stretched in’ all directions. These mesenchymal strands can be seen in figure 43 of the liver of an embryo 20 mm. long. After the embryo reaches a length of from 28 to 30 mm. the hepatic cylinders again begin to increase in size. They rise rapidly in diameter from an average of 35.6 micra in an embryo 28 mm. long to 46 micra in an embryo of 50 mm. and 54 micra in an embryo of 95 mm. This growth is still noticeable in an embryo of 220 mm. in which, however, the cells were too col- lapsed to admit of accurate measurement. While the diameter of the tubules increases, the size of the lumen and the number of cells surrounding it at any given plane as steadily decreases. Thus the average number of cells seen in cross section of a cylinder at 25 mm. is 8, at 50 mm. 4.6, and at 95 mm. 3.7. This indi- cates that the increase in size of the cylinder is due to the growth of the individual cells forming it and not to their multiplication, as is the early increase in size found in embryos from 13 to 16 mm. long. Almost all this growth is due to the deposition of fat in the hepatic cells, the nuclei of which remain stationary or actually decrease in size. While the cells begin to increase in size even when the cylinders are becoming reduced in caliber, due to factors already mentioned, this growth is not sufficient to make up for the reduction until the embryo reaches a length of more than 30 mm. The thick sections illustrated in figures 40, 41 and 42 show graphically some of the changes just described. Figure 40 is a section of a liver in which the process of anastomosis is well under way. It shows the large size and irregular caliber of both the cylinders and their lumina. Figure 41 is from a specimen in which the hepatic sinusoids have nearly reached their highest 264 RICHARD E. SCAMMON development and in which the cylinders are reduced to slender tubes. Figure 42 is from a somewhat older specimen in which the cylinders have again begun to increase in diameter. The changes in the number of cells surrounding the lumen of the cylinder in cross section are shown in the four figures forming plate 3. My remarks’ upon the changes in the finer structure of the hepatic cells can be regarded as little more than notes made in the course of the general study of the growth of the cylinders. During the development of the hepatic tubules and cylinders, the nuclei of. the cells forming them are modified in shape, size and structure. At the time when the tubule ridges first appear upon the hepatic pouches the hepatic nuclei are elongately oval in outline, their longer axes averaging from 14 to 15 micra and their lesser axes from one-third to one-half of this length (fig. 28).* With the definite outpouching of the hepatic tubules the nuclei become broader and shorter, but their volume remains practi- cally unchanged (figs. 29, 30). In an embryo 10 mm. in length (S.C. 20) from which tubules have been described and figured in the preceding part of this paper, the hepatic nuclei average 8 micra in diameter and 11 micra in length. Anastomosis has apparently no effect upon either the size or structure of the nuclei. In embryos from 10 to 20 mm. long one can follow the change in shape of the majority of the nuclei from broadly oval to spheri- cal bodies. In a 20.6 mm. embryo (H.E.C. 1494, fig. 32) the great majority of the nuclei are spherical and have a diameter of 10 micra. Thereafter they gradually decrease in size even as the cells grow in size through an increase in fat content. In an embryo 47.3 mm. long (8.C. 11) the average diameter of the nuclei is 7.5 micra and in one of 95 mm. (H.E.C. 1882) it is little if any less. As the fat accumulates in the liver cells the nuclei may again change in shape, being in many cases pressed against the margin of the cell and assuming an oval or crescentic outline suggesting the form of the nuclei found in true fat cells. The nuclei which remain spherical have an average diameter of 7 micra or a little less (in embryos 200-240 mm. long). 6 All measurements given here were made from paraffin sections. HISTOGENESIS OF THE LIVER 265 In early stages the nucleus ‘is always in the basal portion of the cell. After anastomosis they are found more centrally located. In later stages they are found scattered, sometimes in the center of the cell surrounded with a film of protoplasm from which threads extend to the cell periphery; often close to the lumen wall of the cell and sometimes near some other point in its periphery. Before the tubule anlagen appear the hepatic nuclei contain one, two or three large masses of chromatin which surround nucleoli and are generally applied closely to the nucleus wall. These chromatin masses are round or oval in shape and smooth in outline except for one or two small chromatin threads which may extend outward from each mass. The remaining space of the nucleus is filled with a clear nuclear sap through which run a few delicate and faintly staining fibrils. Such nuclei are characteristic of the young Acanthias embryo being found also in the cells of the mesenchyma, mesothelium, walls of the medul- lary canal and in the mesonephric tubules and duct. Those of the mesenchyma have been fully described by McGill (10) in a study of the development of the striated muscle of the oesophagus in the dogfish. The figures in Neal’s (14) recent work illustrate the similar structure of the nuclei in the nervous system. Figure 28 shows the structure of a number of these nuclei stained with iron-hematoxylin. As the hepatic cells are differentiated the chromatin masses above mentioned become more irregular and there are given off from them a number of chromatin strands which eventually form a coarse network. The chromatin masses are reduced in size and may become detached from the nuclear wall. They also may be somewhat broken up and present a granular appearance. On the other hand, the chromatin threads which have originated from them may fuse forming secondary and generally smaller karyosomes which do not surround nucleoli. These changes are shown in figures 31 to 34, 38 and 39. The adult hepatic nucleus is rich in chromatin which is arranged in a coarse network containing several karyosomes some of which are probably the remains of the original ones and some of which are the result of secondary aggregation of chromatin granules from the chromatin threads. 266 RICHARD E. SCAMMON Most of these karyosomes are applied to the nuclear wall. The hepatic cell retains the typical embryonic nuclear structure much longer than do the cells of the mesenchyma, mesothelium, nervous system or urogenital system. This typical embryonic structure is lost first in the cells forming the gall bladder and major hepatic ducts, next in the minor hepatic ducts which are formed from cylinders already well started upon a development towards typical hepatic parenchyma, and finally from the hepatic cells proper. As has been remarked, fat droplets as indicated by vacuoles in the protoplasm of the hepatic cells appear, when the embryos obtain a length of about 25 mm. The use of special reagents would no doubt demonstrate the presence of fat prior to this stage. The droplets are found at first at the base of the cell, but later, in embryos 65 to 95 mm. long, droplets are found scattered through the entire cell body reducing the protoplasm to the network which has been described for the adult hepatic cell of selachians by Shore and Jones (’89) and Pilliet (’90). In Acanthias embryos about 30 mm. long the gall bladder begins to press against the hepatic tissue which lies on either side and above it. This process is probably brought about by the great growth of the internal yolk-sac which lies below the gall bladder. This pressure of the gall bladder causes some degeneration of the hepatic tissue immediately surrounding it. Toldt and Zuckerkandl (’78) have described a similar process in the human embryo. 5. Development of the hepatic tubules in other forms of selachians The development of the hepatic tubules in Torpedo, Raia and Mustelus differs somewhat from that of Acanthias. In these forms the lateral hepatic pouches do not reach the great development found in Acanthias and tubules are formed from these structures at a comparatively earlier stage. The omphalo- mesenteric veins are somewhat larger than in Acanthias and at the time when the individual hepatic tubules develop veinous channels are found both medial and lateral to the hepatic pouches. HISTOGENESIS OF THE LIVER 267 The tubules which are formed from the dorsal parts of the pouches and of the pars hepatica medialis arise singly or in small clusters as do those in the same situation in Acanthias, but in the forms under discussion the dorsal or marginal tubules form a much greater part of the whole number produced than is the case in Acanthias. In Torpedo and in Mustelus the tubules from the ventral portion of the lateral pouches arise as in Acanthias from tubule ridges. These ridges are but slightly elevated and the corresponding grooves on the internal surfaces of the pouches are wide and shallow. The individual tubules which form from these ridges do not remain as mound-like structures arranged in rows but grow out almost at once and like the dorsal tubulesappear as slender tubes extending outward between the veinous sprouts. In most cases the hepatic tubules which are first formed come in contact with the splanchnic mesothelium covering the liver. The omphalo-mesenteric veins in these forms increase in size very rapidly and before tubule formation has progressed very far the anterior, ventral and lateral surfaces of the liver pouch are practically surrounded by a venous lake. All tubules which arise after this period project into this sinus and in their growth they push its endothelial wall before them. This process is illustrated in figure 44, which is of a thick frontal section of the liver of an embryo of Mustelus canis approximately 12 mm. long. The number of tubules arising directly from the pouch walls is small in these forms as compared with Acanthias for the surface area of the pouches is much reduced from the first. The large majority of the later tubules is formed by the branching of the earlier ones. In Acanthias the hepatic tubules in the earlier period of their development are short, broad and pouch-like and the branches which arise from them are nipple-like projections. In the other forms studied both the primary and secondary tubules are slender elongated tubes which are early united in a complex branching and anastomosing network, the meshes of which are separated by sinusoids of large size. Figure 44 shows an anastomosis of a young hepatic tubule in Mustelus, and figure 45 of a similar 268 RICHARD E. SCAMMON section of an embryo about 4 mm. longer of the same species shows the complete establishment of the hepatic network. While the early tubules of Raia, Mustelus and Torpedo are smaller in cross-section than are those of Acanthias, the lumina of these tubules are considerably larger. The cells lining the lumina are cubical or low columnar in outline as compared with the high columnar type found in Acanthias. The nuclear struc- ture is quite similar to that of the hepatic cells in Acanthias, the chromatin covered nucleoli described for that form being par- ticularly prominent in Torpedo and Mustelus. The later history of the cylinders is quite similar to that of Acanthias. The cylinders rapidly increase in diameter and the contained lumina become smaller. The increase in diameter of the cylinders is due, as in Acanthias, to the growth of the individual cells, and the number of cells about the lumen in any given section steadily decreases. The nuclear changes are similar to those found in Acanthias. In such specimens of Mus- telus and Squatina as J have examined the fat contained in hepatic cells remains in discrete droplets instead of forming one large mass as is generally the case in Acanthias. The same is true to some extent of Torpedo (fig. 35). Pilliet (90) has described areas of comparatively undifferenti- ated cells in the adult selachian liver. These cells form the portions of the cylinders which lie about the hepatic-portal veins. They are of comparatively small size and contain centrally placed nuclei which stain deeply with alum-carmin. The fat content of the cells is less than of those cells located elsewhere, and particularly of those lying in the neighborhood of the hepatic veins. Pilliet regards these smaller cells as reserves or nests of young cells which, from their position near the nourishing vessels, contribute to the growth of the organ. Apparently he did not find them in all the specimens which he examined. I have seen no evidence of a retardment of the differentiation of the cells near the hepatic-portal veins in Acanthias embryos, and if such occurs it must be at a late period in the development of the organ. ‘There are numerous small cells in the immediate neighborhood of the hepatic-portal vessels, but they all form ’ HISTOGENESIS OF THE LIVER 269 portions of the terminal bile ducts, the development of which will be described in a later part of this paper. Such embryos of Torpedo as I have examined agree with Acanthias in lack of any definite nests or reserves of undifferentiated cells about the hepatic-portal veins. Jn well advanced Mustelus embryos, however, the hepatic cells which surround the larger branches of the hepatic-portal veins do remain somewhat smaller than those of the remainder of the liver and are not so completely charged with fat. In summary it may be said that the chief differences between the two types of selachian liver, that represented by Acanthias and that by the several other forms mentioned, lies almost en- tirely in the earlier stages. These differences seem to be depend- ent on the difference in the size and arrangement of the omphalo- mesenteric veins at the time of the formation of the hepatic tubules. DEVELOPMENT OF THE MINOR RAMI OF THE HEPATIC DUCTS The formation of the minor rami of the hepatic ducts is closely associated with the history of the hepatic cylinders. In a pre- ceding publication (Scammon 713) it was stated that all of the major and some of the minor rami of the hepatic ducts arise from the constricted bases of certain fairly definitely placed clusters of hepatic tubules. In following the history of these tubule clusters it was found that they become separated from the hepatic pouch from which they arise by a broad and rather in- definite peduncle which is at first hardly more than an extension of the pouch wall. Later this peduncle becomes constricted and elongated, forming a small branch from the pouch which by this time is transformed into a segment of the hepatic duct. This development begins before tubule anastomosis gets fairly under way and continues at the time when anastomoses are taking place. No new major rami arise as outpouchings after anastomo- sis is established and aside from the actual lengthening of already established rami, which is not great, all further growth of these structures takes place by the transformation of pre-existent hepatic tubules into ducts. 270 RICHARD E. SCAMMON This illustrates well the two methods of hepatic duct formation. The first and more primitive type is that suggested by Minot (93, p. 763) in which the duct is the result of a direct outpouch- ing of the wall of the hepatic diverticulum. In the second and specialized type the duct is produced by the transformation of portions of the network of hepatic cylinders. Selachians clearly stand near the bottom of the scale in this phase of development. Here the ductus choledochus, the hepatic ducts and their major rami arise as outpouchings and only the minor rami and the most distal portions of the major ones are of trabecular origin. In ganoids, amphibians, reptiles and birds the ductus choledochus and the proximal part, at least, of the hepatic ducts are formed by outpouching and the remainder of the duct system from tra- beculae. In the mammals apparently the ductus choledochus alone is the result of an outpouching, but further investigation may change this conception. The differentiation of hepatic cylinders into bile ducts is very closely associated with their relations to the blood vessels. Bile ducts are only formed from cylinders which are in contact with the main trunks of the hepatic-portal veins or their larger and more definite branches. Still more striking is the fact that the side of the cylinder towards the vessels precedes in a very marked degree the differentiation of the opposite side and in fact in the smallest ducts the cells of the opposite side of the cylinder may never be transformed into duct epithelium at all but complete the ordinary development of true hepatic cells. Such a terminal duct from the liver of an embryo 95 mm. in length is shown in figure 37. On the other hand, of the many cylinders which closely surround the vascular trunk only a very small percentage is transformed into ducts. The development of the minor ducts is extremely small in proportion to the amount of hepatic parenchyma, smaller, I think, than in any other group of verte- brates. There is absolutely no indication of any system of intercalated ducts. In the differentiation of a hepatic cylinder into a bile duct the former first approaches more nearly a perfect circle in cross HISTOGENESIS OF THE LIVER PFA section and the lumen distinctly enlarges. The cells on the vascular side of the lumen are reduced in actual height but be- come more columnar in form because of the still greater reduction of the size of their bases. At the same time the nuclei which in hepatic cylinders are round or broadly oval in section and lie near the center of the cells become elongately oval in outline and tend to retreat to the bases of the cells. The cells are so di- minished in size that the nuclei which increase little, if any, in bulk almost fill them. The nuclei lose their typical structure of a clear karyoplasm containing one or two large chromatin masses from which radiate chromatin threads and present in- stead a reticular chromatin network made up of evenly distrib- uted granules of about the same size. The protoplasm becomes homogeneous and colors darkly with plasma stains. An example of such a developing duct at an early stage is shown in figure 39. Approximately one-third of the duct which abuts upon a blood vessel shows considerable progress in differentiation, while the cells of the segment opposite it are true hepatic cells. Between the two are zones of transitional cells. To the side of this duct is a smaller one in still an earlier stage of differentiation. The out- line sketch shown in figure 5 illustrates the changes in shape of the hepatic cells at the time of duct formation. The cells of the large hepatic ducts which are formed from the hepatic pouches and their evaginations show much the same steps in cytomorphosis as do those just described. The nuclei of the larger ducts, however, are oval at the start and so undergo no changes in form, but the change in chromatin arrangernent is the same as in the minor ducts. In the gall bladder the same changes also take place but at a late stage (60-80 mm.) the nuclei again become circular in cross section and come to occupy the centers of the cells which are much elongated. It is sometimes stated that the bile duct epithelium is formed of cells of a more primitive type than those of the cylinders or trabeculae. In the forms under discussion, however, the bile duct cells have departed farther from the embryonic type than the parencyhmal cells, if we may judge by their nuclear structure. 272 RICHARD E. SCAMMON Fig. 5 Minor hepatic duct with connecting hepatic cylinders from the median lobe of the liver of an Acanthias embryo 37 mm. long (H.E.C. 3538). > 400. Lumina of hepatic duct and cylinders solid black; hepatic cells in stippled outline; sinusoids, S, in unbroken outline. DEVELOPMENT OF .THE HEPATIC MESENCHYMA It is well known that the mesenchymal tissue of the selachian liver is extremely scanty in the embryo although fairly abund- ant around the blood vessels in the adult. No particular study has been made of its origin, but Minot (’01) in his study of sinusoids, and more recently Debeyre (’09) in following the de- velopment of the hepatic cylinders, have noted the occasional delamination of mesenchymal cells from the mesothelium cover- ing the liver and the presence of such elements along the walls of the hepatic sinuses. Debeyre also figures a cross section of a small portion of the periphery of the liver of an Acanthias embryo 30 mm. long, which shows small spurs of mesenchymal tissue extending inward from the mesothelium between the hepatic HISTOGENESIS OF THE LIVER 2s cylinders. The origin of the hepatic mesenchyma can be followed with little difficulty in Acanthias. Here it appears that this material is probably derived completely from the mesothelium: but at two distinct periods and from two different regions, and that in both cases this proliferation is associated with distinct irregularities of splanchnic mesothelium. The first mesenchymal proliferation appears at a little later time than the formation of the stroma of that portion of the gut which lies posterior to the liver. In Acanthias embryos 5 mm. in length the liver pouch is but slightly differentiated and con- sists of two shallow diverticula which lie in the lateral walls of the archenteron and are fused anteriorly to form the median hepatic pouch. The archenteron of this region is clothed on either side by a layer of splanchnic mesoderm which is continuous above with the radix mesenterica and below with the splanchnic layer of the blastoderm. At the radix this layer is much thickened but it is reduced to a moderately thin layer over the sides of the archenteron. The irregular endothelial walls of the omphalo-- mesenteric veins intervene between the ventral part of the _archenteron and the mesothelium, but no mesenchyma is present. Irregular processes from the mesothelium, however, do extend inward and in places it appears as though cells were about to be delaminated. A short time later the hepatic pouch increases much in size and extends anteriorly far in front of the anterior intestinal portal. With this change the ventral parts of the investing layers of splanchnic mesothelium are brought in con- tact and eventually fuse, thus forming for some time a ventral mesentery. In connection with this process there appear two distinct sets of mesothelial irregularities. These consist of the mesothelial villi on the right side and of numerous irregular folds on the left. The mesothelial villi found in connection with the covering of the selachian liver were first described by Choronshitzky (’00), although they were observed in other forms long before that time. Hochstetter (00) has given their later history in Acanthias in connection with his study of the formation of the septum transversum. As Hochstetter has stated, these structures are 274. RICHARD E. SCAMMON found only on the right side. They appear in embryos of 45 to 30 segments as a thickened plate of the mesothelium overlying the right omphalo-mesenteric vein (fig. 6A). This plate is soon thrown into a series of pouch like irregularities, and the spaces thus formed are filled with a delicate network of protoplasmic processes from the mesothelial cells. The cores of the villi are thus at no time really empty, and are soon occupied by delaminated mesenchymal cells. From the bases of these villi mesenchymal cells are proliferated apparently both from the walls and the mesenchymal core, and soon come to form a small mass on the right side lying just below the omphalo-mesenteric vein. On the opposite side the irregularities are not villi but longitudinal folds. Unlike villi they do not arise from a thickened plate, the mesothelium at this point remaining as thin as else- where at the time of formation. Afterward the cells become more columnar and a mass of mesenchyma, larger but similar in other respects to that of the opposite side, is proliferated and underlies the left omphalo-mesenteric vein. Figure 7 shows the position and extent of these two mesenchymal proliferations in an embryo 7.5 mm. long (H.E.C. 1496). Soon after this stage the ventral mesentery breaks down and the ventral surface of the liver is free throughout its extent. With this process there is a delamination of mesenchymal cells along the median . ventral line which unites the lateral ones already described, and there is thus formed a general ventral bed of mesenchyma which, as growth proceeds, forms a coating about the gall bladder and constitutes a large loose-meshed mass which extends forward from the gall bladder to the anterior mesothelial wall of the liver. This constitutes the first and ventral contribution of mesen- chyma to the liver from its mesothelial envelope. In its entire extent it produced the mesenchyma which surrounds the gall bladder, and later the vessels adjacent to it, the covering of the cystic duct, and to an undeterminable degree the sparse mesen- chymal tissue of the lower part of the hepatic parenchyma. The irregularities of the mesothelium on the right side do not | continue for any great period. By the time the embryo reaches HISTOGENESIS OF THE LIVER Dae Fig. 6 400. 304 PLATE 2 HISTOGENESIS OF THE LIVER RICHARD E. SCAMMON 28 PLATE 3 EXPLANATION OF FIGURES 31 to 34 Four sections of livers of Acanthias embryos of different ages, illus- trating the changes in the hepatic cylinders; all * 400. 31 Section of the liver of an embryo 16 mm. long (K.U.E.C. 547); iron hematoxylin. 32 Section of the liver of an embryo 20.6 mm. long (H.E.C. 1494); iron hematoxylin. 33 Section of the liver of an embryo 32 mm. long (H.E.C. 1652); iron hematoxylin. 34 Section of the liver of an embryo (of Squalus suckli1) 95 mm. long (H.E.C. 1882); alum cochineal. 306 PLATE 3 HISTOGENESIS OF THE LIVER RICHARD E. SCAMMON es 6) noes Se Ne WN < If. oe \ ga 7 jy Weg Y (eas ) SES < By, \e y, \ Ney €) as 32 MG a Se Nay: Dees Nee ae ae ee tae Oe SC ie @ gy Se 2 Oe Por oi oS ) / : eS \ SSN ee) ay ee Cor I @ (@ ae Ca? Os ee? | € On fe A. EN a Py) Nh, oe me ; aa. Y= ve 4 es LL @ be Gre Ya oe: < ave! & | Be ~S ef 1 yee / ey “ 3) LE f_/ (Ad IN ee Sa | [Y \ a i y \y a ee 3! 2 : ae f Mey) See ie LT oo we oS 7 sot ae Co OO SS ee “ | tn | Gs XCD | 34 307 33 PLATE 4 EXPLANATION OF FIGURES 35 Transverse section of the margin of the liver of an embryo of Torpedo ocellata 29 mm. long; alum hematoxylin. 350. 36 Section of an anastomosis between two hepatic tubules of an Acanthias embryo 14 mm. long (S.C. 30); alum hematoxylin. »X 500. (This section shows a stage intermediate to those shown in figures 26 and 27). 37 Transverse section of a terminal bile duct of an embryo 95 mm. long (H.E. C. 1882); alum cochineal. X 400. 38 Hepatic tubules and mesenchymal ingrowths from a transverse section of the dorsal margin of the liver of an Acanthias embryo 24.7 mm. long (H.E.C. 1492); iron hematoxylin. »X 400. M, mesenchymal ingrowths. H.t., hepatic tubules. 39 Transverse section of a developing bile duct of an Acanthias embryo 24.7 mm. long (H.E.C. 1492); iron hematoxylin. X 400. 308 PLATE 4 HISTOGENESIS OF THE LIVER RICHARD E. SCAMMON exis oes “ie Ko , LE \ \ may y af f 374 rose ; aay : : y fer * R < a vad Pa Cares GQ) yes 39 309 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No.3 PLATE 5 EXPLANATION OF FIGURES 40 to 42 Portions of thick sections of livers of Acanthias embryos of dif- ferent ages. The method of preparation is described in footnote 4, page 225. All approximately 200. 40 Thick section of the liver of an embryo approximately 15 mm. long. 41 Thick section of the liver of anembryo approximately 21 mm. long. 42 Thick section of the liver of an embryo approximately 30 mm. long. 3510 PLATE 5 HISTOGENESIS OF THE LIVER RICHARD E. SCAMMON 311 PLATE 6 EXPLANATION OF FIGURES 43 Preparation of the periphery of the posterior lobe of an Acanthias embryo approximately 20 mm. long showing the terminal buds of the hepatic network and the attachment of the network to the mesothelial covering by means of mesen- chymal strands. 44 Thick frontal section of the liver of an embryo of Mustelus canis approxi- mately 12 mm. long. The large hepatic duct and the hepatic tubules lie in a vascular sinus which is represented in black. The vertical band seen on the right hand side is a section of the mesothelial covering of the liver. 312 HISTOGENESIS OF THE LIVER PLATE 6 RICHARD E. SCAMMON 313 PEATE 4 EXPLANATION OF FIGURE 45 Thick frontal section of the liver of an embryo of Mustelus canis approxi- mately 16 mm. long. : 314 HISTOGENESIS OF THE LIVER RICHARD E. SCAMMON PLATE 7 315 i i { i i: THE DEVELOPMENT OF THE THYMUS IN THE PIG I. MORPHOGENESIS J. A. BADERTSCHER From the Department of Histology and Embryology, Cornell University, Ithaca, N. Y. TWELVE FIGURES—TWO PLATES HISTORICAL It has been definitely established by many investigators that the thymus of all mammals is of epithelial origin. More recent investigations have shown, however, that the epithelial anlage of the thymus is not derived from the same germ layer in all mammals. Investigators are agreed on the point that the thy- mus of mammals, when it is of a purely entodermal origin, is a, derivative of the ventrally directed epithelial diverticulum of the third pharyngeal pouch. It has also been quite definitely settled that in some mammals (mouse, Roud ’00) the thymus is entirely of ectodermal origin. The mixed (ectodermal-ento- dermal) origin of the thymus in some mammals has not yet been generally accepted. In pig embryos it is the close topographi- eal relation that exists between the cervical vesicle and the third pharyngeal pouch that makes a mixed origin of the thymus possible. Among some of the workers on the early development of the thymus of the pig may be mentioned Fischelis, Kastschenko, Zotterman, Born, Bell, and Fox, the first three of whom attribute to the thymus an ectodermal-entodermal origin. Fischelis (’85) derived the thymus from the third pharyngeal pouch and the third branchial groove. According to this investigation these two fuse, and from their point of fusion each contributes about one-half to a ventrally directed downgrowth, the anlage of the thynrus. This conclusion is erroneous, for that portion of the 317 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 3 318 J. A. BADERTSCHER thymus which is derived from the third pharyngeal pouch is a comparatively long mid-ventrally directed epithelial tube before the cervical sinus is fused to it. He also makes no mention of the XII cranial nerve which plays an important part in modi- fying the topographical relations of the anterior portion of the thymus to surrounding structures. Basing his conclusions on inaccurate observations, his views in regard to a mixed origin of the thymus have now only an historical value. The first detailed study of the early development of the thymus in the pig was made by Kastschenko (’87). He describes the mesial portion of the sinus cervicalis, which he calls the ‘vesicula thymica,’ as fusing with the anterior end of the epithelial anlage of the thymus, which is derived from the third pharyngeal pouch. In the shiftings of some of the structures in the neck, that occur in young embryos during growth, the lateral portion of the cervical vesicle is separated mechanically from its mesial portion by the hypoglossal nerve. The free lateral portion of the cervical vesicle gives rise to the ‘thymus superficialis’ which is necessarily of ectodermal origin. He claims that the superficial thymus is not a constant structure, for, in a 30 mm. embryo that he examined, it was not present. The anterior end of the thymus to which the mesial portion of sinus cervicalis has fused, plus the parathyroid, that lies close to it, he designates the thymus head; while the large remaining por- tion of the thymus which is of a purely entodermal origin, plus the thymus head, he calls the ‘thymus profunda.’ The largest embryo examined by him was 82 mm. in length. Zotterman (711) also made a detailed study of the morpho- _ genesis of the thymus of the pig. Her conclusions are in accord with those of Kastschenko with the exception that in about one-half of the specimens examined the superficial thymus was connected with the thymus head by a cord of cells that looped over the hypoglossal nerve. The superficial thymus was found in all the specimens examined and in the largest (105 mm.) investigated all the features common to the thymus (cortex, medulla, Hassall’s corpuscles, etc.) were present. Fox (’08) agrees with Kastschenko that the superficial thymus DEVELOPMENT OF THE THYMUS 319 of the pig arises by a constriction of the fundus praecervicalis, but claims that in embryos up to 35mm. in length, the oldest stage he examined, its histological structure does not resemble that of the thymus more than any other branching epithelial mass. Born (’83) derived the thymus anlage in pig embryos from the third pharyngeal pouch, while Bell (’06) also is inclined to believe that the ectoderm takes no part in the formation of the thymus. MATERIAL AND METHODS For investigation of the early stages of the morphogenesis of the thymus, the excellent collection of pig embryo series, 3 to 42 mm. in length, in the Department of Histology and Embry- ology of Cornell University, proved very helpful. In addition to these, five embryos ranging from 9 to 21.5 mm. in length, and the neck and upper thoracic region of eight embryos ranging from 32 to 95 mm. in length, were sectioned transversely. These sections, 10 microns in thickness, were stained with hematoxylin and eosin. From series of this group reconstructions of the pharyngeal region were made. Many dissections exposing the thymus were made of the neck and upper thoracic region of embryos from 100 to 280 mm. in length (full term). The thy- mus of a pig one day old was also examined. MORPHOGENESIS In the investigation on the morphogenesis of the thymus special _ attention was constantly directed toward the development of the superficial thymus because its existence is not yet generally accepted and, since the latest developmental stage in which it was investigated by Zotterman was only 105 mm. in length, its fate is not definitely known. An 11 mm. embryo was the developmental stage chosen as the starting point for the study of the morphogenesis of the thymus. At this stage the ectodermal and entodermal parts of the branchial grooves and pharyngeal pouches can still be 320 J. A. BADERTSCHER distinguished from each other without difficulty. The sinus cervicalis, formed by the rapid growth in a caudal direction of the mandibular and hyoid arches and the more retarded growth of the branchial arches proper, is already well mapped out. As this stage shows well accepted relations and developmental steps it needs no further description. Embryo of 14.5 mm. (figures 1 and 2). This is the youngest developmental stage from which a reconstruction of one side of the neck was made. Since the determination of the real origin of the thymus was one of the crucial points in this inves- tigation only that part of the neck containing the anterior por- tion of the thymus.anlage and the vesicula cervicalis was modelled. The posterior edge of the hyoid arch has grown over the open- ing of the sinus cervicalis, shutting it off from the exterior. The cavity thus formed is the vesicula cervicalis (V.c.) or the ‘vesicula thymica’ of Kastschenko.! The vesicula cervicalis, now widely separated from the ectoderm (Kci.), is still connected with it by a heavy cord of cells, the ductus cervicalis (D.c.). Only in places through its entire extent are traces of a lumen left. To the outer end of the vesicula cervicalis is attached a cord of cells that runs in an antero-ventral and mesial direction and con- nects with the second pharyngeal pouch. This is the ductus branchialis (D.b.).. The anterior one-fifth of this structure pos- sesses a lumen which is continuous with that of the second pouch. At this stage it is impossible to determine the extent of the part that is of entodermal origin and the extent that is of ectodermal origin. The boundary line between the two has disappeared through the obliteration of the lumen. Fox (’08) was unable to find the ductus branchialis in pig embryos but demonstrated a long diverticulum—‘filiform process’—arising from the ventro- lateral angle of the second pouch and connected with the ecto- 1 “According to H. Rabl (’09) the term ‘vesicula cervicalis’ is to be applied to the entire complex, including the two ductus branchiales; Hammar uses the term ‘vesicula praecervicalis’ only for the vesicular portion that is associated with the third pharyngeal pouch, this portion being approximately identical with the ‘fundus praecervicalis’ (cervicalis) of His and H. Rabl, as well as with the ‘vesicula thymica’ of Kastschenko and the ‘sinus vesicle’ of Zuckerhandl.’’ (Quoted from Keibel and Mall’s Human embryology, vol. 2, p. 456). DEVELOPMENT OF THE THYMUS 5 Pa! derm. This process was not present in any of the embryos I examined. The vesicula cervicalis (fig. 2, V.c.) lies lateral to the third pharyngeal pouch, between the cephalo-dorsal portion of the parathyroid (Pt. 3) and the caudo-ventral part of the ganglion nodosum (G.n.) and a short distance anterior to the hypoglossal nerve (N.XJI). Its general shape is fusiform, with its long axis almost perpendicular to the surface ectoderm. ‘The middle third is solid, but each extremity contains a cavity. The expan- sion of the central portion is due to a proliferation of the cells of its anterior wall which presses tightly against the dorsal sur- face of the parathyroid. The pressure against the parathyroid has apparently caused the obliteration of the cavity of the vesicula cervicalis in its central portion. Its inner third lies closely along the ventral side of the ganglion nodosum into which its curved end projects and with which it apparently is fused. The parathyroid (Pt. 3) is now a massive structure lying lateral to the third pharyngeal pouch (S.b.3), anterior to the hypoglossal nerve with which it is in contact, ventral to the ganglion nodosum and dorsal to the carotid artery (A.c.). Its general shape is that of a hemisphere with its flat side turned toward the vagus nerve and the vesicula cervicalis. A depres- sion in both the vesicula cervicalis and the parathyroid mark the points of most intimate contact between the two structures. The entodermal anlage of the thymus (7.e.) is now, through- out its greatest extent, a solid cord of cells, and still attached to the third pharyngeal pouch. Its anterior end lies closely against the parathyroid and the hypoglossal nerve. From its point of origin it extends caudally, and, with the exception of about one- fourth of the posterior portion, lies ventro-laterally to the carotid artery and the vagus nerve. ‘The caudal portion makes a rather sharp turn in a ventro-mesial direction and occupies the upper part of the pericardial region. The diameter of the caudal part is considerably greater than that of the remaining portion. This is due to the presence of a large lumen and thick walls in this region. In the central and anterior portions only a few slight traces of a lumen persist. The vesicula cervicalis and the an- 322 J. A. BADERTSCHER terior portion of the entodermal thymus do not come in contact with each other in this developmental stage. Embryo of 17.5 mm. (figures 3 and 4). During the interval between this and the previous stage shiftings in the pharyngeal region have taken place that have changed the relation of some of the parts to each other. The ductus cervicalis (D.c.), now a solid cord of cells, is still connected with the ectoderm. The ductus branchialis on the left side, which was not modelled, has lost its connection with the outer end of the vesicula cervicalis. No traces of it in this region can be seen. It is, however, still connected with the second branchial pouch from which it extends for a short distance toward the point of its former attachment.? On the right side it is still a continuous solid cord of cells: ex- tending from the second pouch to the vesicula cervicalis. In a 21 mm. pig embryo Zotterman (11) had demonstrated the ductus branchialis as a continuous cord of cells while the ductus cervi- calis as being broken. In the embryos which I examined the ductus branchialis was always the first to become discontinuous. The vesicula cervicalis (V.c.) no longer lies perpendicular to the ectoderm. The vesicula cervicalis medialis (V.c.m.)? extends from its point of attachment to the ganglion nodosum in an antero-lateral direction to the hypoglossal nerve (V.XIJ) around which it forms an acute angle. Frem the nerve the vesicula cervicalis lateralis (v.c.l.) extends for a short distance in a caudo- lateral direction. This is the same general direction taken by the ductus cervicalis which is connected to the vesicula cervi- calis lateralis and the ectoderm. ‘The vesicula cervicalis medialis is tightly wedged in between the vagus on its dorsal side and the parathyroid gland and a small portion of the thymus on its 2 A reconstruction of this remnant was deemed unnecessary since it takes no part in the formation of the thymus and would have needlessly increased the size of the model. 3 From the reconstruction as represented in figures 3 and 4 it will be seen that the vesicula cervicalis now loops over the hypoglossal nerve. For the sake of simplicity as well as for clearness, that portion of the vesicula cervicalis lying between the nerve and the pharynx will be termed the ‘vesicula cervicalis medialis’ while the part lying between the nerve and the surface ectoderm will be termed the ‘vesicula cervicalis lateralis.’ ok pe ee DEVELOPMENT OF THE THYMUS 32a ventral aspect. Its caudal portion is greatly flattened but as it approaches the nerve it gradually assumes a cylindrical form which also is the shape of the vesicula cervicalis lateralis. A part of its flattened caudal portion dips into the ganglion nodosum while a portion lies in close contact with the thymus. Fusion between the thymus and the vesicle has apparently not yet taken place, for the boundary of both can still be clearly deter- mined. The lumen of the vesicula cervicalis is for the most part obliterated. Only slight traces here and there in its course persist. It is largest in the portion that dips into the ganglion nodosum. Here the lumen is large and the wall of this portion of the vesicle is no thicker than that of earlier stages. Appar- ently no cell proliferation takes place in this region. The sur- face of the entire vesicula cervicalis is more or less irregular. An idea of its shape can best be obtained by referring to figure 4 in which the hypoglossal nerve and a part of the ganglion nodosum were removed, thus almost entirely exposing it. The parathyroid (Pt. 3) is an elongated and very irregular mass of cells that is tightly packed in between the cervical vesicle and vagus nerve on its dorsal aspect, and the carotid artery on its mesial surface. Its caudo-mesial and caudo-lateral portions are in contact with the thymus while its anterior portion is on a lével with the arch of the vesicula cervicalis over the hypoglossal nerve. The thymus (7..e.) is considerably longer than in the preceding stage. Its cephalic and caudal ends have about the same relative position to the other structures as in the 14.5 mm. embryo. Its greater length at this stage is due to growth which has kept pace with the growth of the pharynx. It is still connected with the third pharyngeal pouch by a greatly attenuated cord of cells. Its anterior portion (figs. 3-4) is fused to the caudal aspect of the parathyroid from which it extends caudally. As in the pre- ceding stage, the caudal portion makes a sharp turn in a ventro- mesial direction and lies over the upper portion of the peri- cardium. The caudal portion of the right thymus extends across the mesial plane while the same region of the left thymus lies to the left of the mesial plane and extends farther caudally than 324 J. A. BADERTSCHER the right one. The anterior and central portions are cylindrical in outline, having an almost uniform diameter. In the left thymus the lumen of the anterior and central regions has entirely disappeared, while in the right thymus only a trace of it persists in the central portion. The caudal portion in the pericardial region is greatly enlarged. The lumen in this region is broken but in places is quite large in diameter. The walls are very thick and irregular, no longer retaining their cylindrical shape. The anterior portion of the thymus also extends for a short distance along the ventro-lateral aspect of the parathyroid. It thus has two prongs between which lies the epithelial body. This condition is not present on the right side and was not observed in other specimens of about the same developmental stage. In stages earlier than this the parathyroids are anterior to the hypoglossal nerve. The anterior portion of the thymus is in close contact with both, as shown in figure 2. In the shiftings that occurred during the interval between this and the previous stage it appears that a portion of the thymus was carried along by the nerve and strung along the parathyroid thus bringing about the split condition of its anterior end. Embryo of 21.5 mm. (figure 5). In this stage the vesicula cervicalis has lost its connection with the ectoderm. The duc- tus cervicalis has entirely disappeared. The vesicula cervicalis lateralis (7.s.-V.c.l.) lies lateral to the hypoglossal nerve (N.XIT). It is a large fusiform shaped mass of cells containing in its anterior portion a narrow tortuous lumen. This structure is of a purely ectodermal origin and represents the ‘thymus superficialis’ of Kastschenko. It is connected to the vesicula cervicalis medialis by the pars intermedia or connecting band (P.2.) that loops over the hypoglossal nerve. This band was not observed by Kastschenko, hence he held that the superficial thymus remained free from the remaining portion of the thymus. The vesicula cervicalis medialis, also greatly expanded, has lost its connection with the ganglion nodosum, possesses no lumen, and lies along the antero-lateral side of the massive para- thyroid (Pi. 3) where it is fused with the anterior portion of the thymus. DEVELOPMENT OF THE THYMUS 325 The anterior portion of the thymus (7'.e.) has lost its connec- tion with the pharynx and lies on the dorso-lateral side of the parathyroid and is fused with the vesicula cervicalis medialis. In the region of the fused portion it contains a cavity of con- siderable size while the remaining portion along the epithelial body is without a lumen. From the epithelial body the thymus extends in a caudal and a slightly medial and ven- tral direction as a solid cord of cells. Just anterior to its entrance into the thoracic cavity it is slightly enlarged. The extreme caudal portion which lies within the thoracic cavity turns abruptly in a ventral direction, is greatly flattened, and in contact with the pericardium. The thoracic segments of the right and left thymus at this stage lie closely together but are not fused. Bell, however, in a 20 mm. embryo, describes them as being fused. The hypoglossal nerve now forms an acute angle with that portion of the vagus lying immediately posterior to it. In the two preceding stages that were modelled, the corresponding angle formed by these two nerves was obtuse instead of acute. This change in the form of angle between the earlier and later stages apparently is due to shiftings—a consideration of which is to follow—that take place in the neck during the growth of young embryos by which a stress appears to be exerted on the hypoglossal nerve by the cervical vesicle. The thymus at this early stage (21.5 mm.) can be divided into seven regions, most of which in the later stages become very pronounced. They are: (1) The ‘superficial thymus’ which rep- resents the vesicula cervicalis lateralis and is of a purely ecto- dermal origin; (2) the ‘thymus head’ which represents the struc- ture formed by the fusion of the vesicula cervicalis medialis and the anterior portion of the entodermal anlage of the thymus; (3) the ‘connecting band’ which loops over the hypoglossal nerve and connects the superficial thymus with the thymus head and is of a purely ectodermal origin; (4) the ‘mid-cervical segment’ which is an enlargement of the thymus between the intermediary and cervico-thoracie cords; (5) the ‘intermediary cord’ which connects the thymus head with the mid-cervical segment; (6) 326 J. A. BADERTSCHER the ‘thoracic segment’ which lies in the anterior portion of the thorax and is spread over a portion of the pericardium; and (7) the ‘cervico-thoracic cord’ which unites the mid-cervical segment to the thoracic segment. This system of nomenclature, for which we are indebted to Kastschenko, Zotterman, and Bell, will be used in the discussion of all the later developmental stages. The different regions of the thymus described in the 21.5 mm. embryo were examined microscopically in the following five stages. These will be briefly described in order to present a more complete developmental history up to a 95 mm. embryo, which was the oldest stage in which a part of the pharyngeal region was reconstructed. Embryo of 26 mm. The thymus as a whole is considerably larger than in the preceding stage. The surface of the super- ficial thymus, the thymus head, and the thoracic segment has become very irregular due to outgrowths of epithelial buds from the main stem. This budding represents the beginning of lobu- lation and had already started in the preceding stage. Lobula- tion of the mid-cervical segment has just begun. ‘The super- ficial thymus extends only slightly farther caudally from the hypoglossal nerve than in the preceding stage. The connecting band on the right side has disappeared but the superficial thymus has retained its usual topographical relation to the thymus head. From the parathyroid body the general direction of the thymus is in a caudo-mesia land ventral direction. The intermediary cord on the right side is only a very slender cord of cells while that of the left side has a considerably greater diameter. The cervico-thoraciec cords are short and have a uniform diameter of small dimension. The thoracic segments lie in contact with the anterior and ventral portion of the pericardium. The two segments lie close together and have fused in some places along their median sides. Embryo of 832 mm. The connecting band on the left side is broken. No traces of it can be seen in connection with the thymus head but a remnant of it is still attached to the super- DEVELOPMENT OF THE THYMUS 327 ficial thymus and extends as a greatly attenuated cord of cells toward the hypoglossal nerve. The general features of the entire thymus at this stage so closely resemble those of the 26 mm. embryo that a detailed description is unnecessary. The only difference of importance is the greater size of the organ as a whole and of the epithelial buds from the main stem. Embryo of 40 mm. The connecting band is continuous around the hypoglossal nerve on both sides. The parathyroids are slightly elongated and lie along the dorso-mesial side of the central third of the thymus head. Many of the primary epi- thelial buds of the enlarged segments (the superficial thymus, thymus heads, mid-cervical and thoracic segments) have sent out processes, thus marking the beginning of secondary lobula- tion. The intermediary cords are greatly attenuated and show no signs of budding. The transition from the thymus head to the intermediary cords and from the latter to the mid-cervical segment is very abrupt. The cervico-thoracic cords are short and lie near each other a short distance ventral to the trachea. The thoracic segments of both the right and left thymus are now fused along the greater part of their median plane. They are a little larger than those in the previous stage and have the same general position over the anterior and ventral portion of the pericardium. Embryo of 52 mm. The connecting band on each side loops over the hypoglossal nerve and connects the thymus head with the superficial thymus. The one on the left side is a compara- tively large and irregularly modelled cord of differentiated thymic tissue while the one on the right side is a slender and greatly attenuated cord of epithelial cells. The intermediary cords are still greatly attenuated cords of epithelial cells but are now studded here and there with small epithelial buds. The left cervico-thoracic cord is still slender with a nearly uniform diam- eter while the right one is much larger and has undergone lobu- lation. Both are now of differentiated thymic structure. The enlarged segments of the thymus are appreciably larger than those in the 40 mm. embryo. They have undergone extended 328 J. A. BADERTSCHER secondary lobulation the lobes of which on account of their large size lie in general more closely together than those in the previous stage. The parathyroid lies partly imbedded in the dorso-mesial aspect of the thymus head slightly anterior to its central portion. The superficial thymus lies closely along the antero-lateral aspect of the thymus head but is not fused to it. Embryo of 63 mm. The connecting band on both the right and left sides loops over the hypoglossal nerve. They are com- paratively large and have an irregular surface similar to the left connecting band in the preceding stage. Aside from their greater size the large segments of the thymus in this stage present no striking morphological changes from those of the 52 mm. embryo. Embryo of 95 mm. (figures 6 and 7). This is the oldest stage in which the anterior portion of the thymus was modelled. The left thymus was chosen for reconstruction although the right one would have done equally well. Figure 6 represents a lateral aspect of the thymus head (C.t.) and the superficial thymus (T.s.) The thymus head lies alongside the common carotid artery (A.c.), its anterior end lying near the bifurcation into the external and internal carotid arteries. The parathyroid (Pé. 3) lies about midway between the two ends of the thymus head along its dorsal border and is closely attached to it. The super- ficial thymus (7’.s.) lies Along the anterior half of the lateral aspect of the thymus head. Its dorsal and ventral borders are almost parallel to each other. The caudal border is rounding while its anterior part tapers irregularly into the slender connect- ing band (P.2.) which loops over the hypoglossal nerve (V.XJJ) -and is connected with the thymus head. Figure 7 represents a ventral aspect of the same structures as seen in figure 6. The superficial thymus (7’.s.) and the thymus head (C.t.) are flat- tened laterally. The anterior portion of the thymus head lies in contact with the dorsal border of the hypoglossal nerve. The superficial thymus les closely against the thymus head but is not fused with it. Its posterior border on the right side gradu- ally tapers down to a thin edge in contrast to the blunt posterior border of the left superficial thymus. The diameter of the inter- =—.¢ Sty = IES S Ces 5 DEVELOPMENT OF THE THYMUS 329 mediary cords is considerably greater than in the preceding stages. Lobules are now present along their entire extent. They are a little shorter in this stage than in the 63 mm. embryo while the mid-cervical segment is somewhat longer. The cervico-tho- racic cords are short and lie closely together. The thoracic seg- ments are thin and flat and spread out over the pericardium to the left of the median line. They extend only a short distance to the right beyond the median line. Embryos of 105, 140, 170 and 280 mm., and pig 1 day post partem (text figures A, B, C, D and EH, respectively). In these stages the structures covering the thymus were removed and the entire organ on the left side, undisturbed, was exposed to view. Diagrammatic drawings, representing accurately the out- line of the lateral aspect of the different regions of the thymus, were made. In all cases specimens were selected in which the connecting band on both sides looped over the hypoglossal nerve and connected the superficial thymus with the thymus head. By referring to the figures cited above it will be seen that the comparative size of the superficial thymus (7’.s.) and the thymus head (C.é.) vary somewhat in different developmental stages. In general, the proportional size of the former to the latter is greater in earlier than in later developmental stages. From numerous dissections that were made it was found that the comparative sizes of the two structures vary considerably in embryos of about the same developmental stage or even in those of the same litter. Figure A represents about the average com- parative size of the superficial thymus and the thymus head in embryos of about 105 mm. in length, while the size of the super- ficial thymus of an 140 mm. embryo as represented in figure B is considerably larger than it ordinarily occurs in corresponding developmental stages. Variations in size of the superficial thymi in the same embryo also occur; e.g., the right one in a 170 mm. embryo was an oblong flap that covered the anterior one-fourth of the lateral surface of the thymus head, while the left one is much smaller as represented in figure C. Jn all the embryos examined the superficial thymus was always closely associated with the thymus head but never fused with it. 330 J. A. BADERTSCHER Text figs. A, B, C, D Outline drawings of the exposed left thymus of embryos respectively 105, 140, 170 and 280 mm. (full term) in length; natural size. C.ct., cervico-thoracic cord; C.i., intermediary cord; C.t., caput thymus = thymus head; N.XII, hypoglossal nerve; S.m., mid-cervical segment; S.th., thoracic seg- ment; 7’.s., thymus superficialis. Text fig. E Outline drawing of the exposed left thymus of a ‘runty’ pig, one day old and only 240 mm. in length; natural size. The thymus in this speci- men was a few millimeters shorter than that in the full-term embryo; this is perhaps due to the fact that the specimen was a ‘runt’. C.ct., cervico-thoracic cord; C.i., intermediary cord; C.t., caput thymus = thymus head; N.XIJJ, hypoglossal nerve; S.m., mid-cervical segment; S.th., thoracic segment; T.s., thymus superficialis. DEVELOPMENT OF THE THYMUS Sal From the thymus head the intermediary cord (C.7.) and the mid-cervical segment (S.m.) extend in a meso-ventral direction to the anterior aperture of the thorax ventral to the trachea. In comparatively early stages they are more or less tortuous as represented in figure A, while in later stages their course is nearly straight. The mid-cervical segment in early stages is short and lies immediately anterior to the thorax while the inter- mediary cord is comparatively long as represented in figure A. As development proceeds the mid-cervical segment gradually becomes longer while the intermediary fe becomes shorter as represented in the figures. The cervico-thoracic segment (C.ct.) in all stages is short and of a comparatively small diameter. It lies in the extreme ven- tral portion of the anterior aperture of the thorax. In early stages the cords of the right and left thymus lie closely together and in later stages they fuse with each other. The thoracic segment (S.th.) in later developmental stages is composed of the thoracic portions of both the right and left thymus which have fused in this region. In embryos about 105 mm. in length, and later stages, it is spread over the antero- ventral surface of the left side of the pericardium. The swinging of the right segment toward the left side is already noticeable ina42mm.embryo. This segment is thickest along the median line (3.5 to 4 mm. in full term embryos) and gradually tapers down to a thin irregular edge. The connecting band was present on both sides in the majority of specimens examined. It may, however, be absent either on one or on both sides. Its rupture is apparently due to the growth in length of the thymus not keeping pace with the growth in length of the neck. The expanded caudal portion of the thymus being firmly anchored in the anterior portion of the thoracic cavity, on account of the unequal rate of growth between the neck and the thymus, will exert a pull on that portion of the organ in the neck and thus greatly attenuate or tear the connecting band. Also there is thus a stress exerted on the hypoglossal nerve which apparently tends slightly to change its direction, as stated in the description of the thymus in 21.5 mm. embryo (p. 325). aoe J. A. BADERTSCHER A microscopical examination was made of the superficial thy- mus in various developmental stages, including that of two full- term embryos. Jt was found that the histogenetic processes of this segment kept pace with those in the segments of the thymus which have a purely entodermal origin. CONCLUSIONS The thymus of the pig has an ectodermal-entodermal origin. The respective origin of each segment is as follows: 1. The superficial thymus, which is a derivative of the cervical vesicle, has a purely ectodermal origin. It is a constant struc- ture and, therefore, forms an integral part of the organ. 2. The connecting band is also a derivative of the cervical vesicle and has, therefore, a purely ectodermal origin. In the majority of embryos it persists to birth but may be absent either on one or on both sides. 3. The thymus head, in which is lodged the parathyroid III, is formed by a fusion of a portion of the cervical vesicle to the anterior end of the epithelial diverticulum derived from the third pharyngeal pouch. It has, therefore, an ectodermal-entodermal origin. 4. The intermediary and cervico-thoracic cords, and the mid- cervical and thoracic segments are derived wholly from the epi- thelial diverticulum of the third pharyngeal pouch and have, therefore, a purely entodermal origin. I wish to thank Dr. B. F. Kingsbury for the aid given me in this work. I am also indebted to Dr. David Marine, of the Western Reserve University, for sending me many formalin- preserved embryos of various sizes from which most of the draw- ings of the exposed thymus were made. DEVELOPMENT OF THE THYMUS aoe LITERATURE CITED Bett, E. T. 1906 The development of the thymus. Amer. Jour. Anat., vol. 5. Born, C. 1883 Uber die Derivate der embryonalen Schlundbogen und Schlund- spalten bei Siugetieren. Arch. f. mikr. Anat., Bd. 22. Fiscuenis, P. 1885 Beitrige zur Kenntnis der Entwickelungsgeschichte der Gl. thyreoidea und Gl. thymus. Arch. f. mikr. Anat., Bd. 25. Fox, H. 1908 The pharyngeal pouches and their derivatives in the Mammalia. Am. Jour. Anat., vol. 8. IXASTSCHENKO, N. 1887 Das Schicksal der embryonalen Schlundspalten bei Siugetieren. Arch. f. mikr. Anat., Bd. 30. Roup, A. 1900 Contribution a l'étude de l’origine et de |’évolution de la thy- roide laterale et due thymus chez le campagnol. Bull. Soc. vaudoise des Sc. natur., T. 36. ZOTTERMAN, A. 1911 Die Schweinthymus als eine Thymus ecto-entodermalis. Anat. Anz., Bd. 38. THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, NO. 5 ABBREVIATIONS A.c., carotid artery Pt.3, parathyroid derived from third C.t., caput thymus = thymus head pharyngeal pouch D.b., ductus branchialis S.b. 2, sacculus branchialis II = second D.c., ductus cervicalis pharyngeal pouch Ect., ectoderm S.b.3, sacculus branchialis III = third G.n., ganglion nodosum pharyngeal pouch G.s.c., superior cervical ganglion S.b.4, sacculus branchialis 1V = fourth N:X., vagus nerve , pharyngeal pouch N.XII., hypoglossal nerve T.e., entodermal thymus P., pharynx T.s., thymus superficialis Pi., connecting band V.c., vesicula cervicalis V.c.l., vesicula cervicalis lateralis V.c.m., vesicula cervicalis medialis PLATE 1 EXPLANATION OF FIGURES Figures 1 to 7 were drawn by Miss Cora Whitman, from wax models which were made by the author. The text figures A to H were drawn by the author. 1 Drawing of a reconstruction of the pharynx and derivatives of the second and third pharyngeal pouches of the right side, including portions of the struc- tures closely associated with the pharyngeal derivatives and a portion of the ectoderm. Ventral aspect; pig embryo 14.5 mm. in length. Model x 92, reduced one-half. 2 Drawing of a reconstruction of a portion of the pharynx, the third pharyn- geal pouch, anterior portion of the thymus anlage, parathyroid 3, cervical vesicle, and associates of the above named structures. Caudo-ventral aspect. The model represented in this figure was made from the same side of the same em- bryo (14.5 mm.) from which the model represented in figure 1 was made. Model 245, reduced one-half. 3 Drawing of a reconstruction of the same structures as enumerated under figure 2. This model was made to show specially the relation of the cervical vesicle to the thymus anlage and the hypoglossal nerve after the shifting of the structures in the neck have become quite noticeable. Ventral aspect, left side; pig embryo 17.5 mm. in length. Model < 182, reduced one-half. 334 DEVELOPMENT OF THE THYMUS PLATE | J. A. BADERTSCHER PLATE 2 EXPLANATION OF FIGURES 4 Drawing of the same model as represented in figure 3, with the hypoglossal nerve and a portion of the ganglion nodosum removed to expose the cervical vesicle and more clearly to show its relation to the thymus anlage. Dorso-lateral aspect; reduced one-half. 5 Drawing of a reconstruction showing the relation of the vesicula cervicalis lateralis (7'.s.) to the thymus head (C.t.) and their topographical relation to neighboring structures. In this stage the thymus and the cervical vesicle, which have fused, have lost their connection respectively with the entoderm and ecto- derm. Right side, lateral aspect; pig 21.5 mm. in length. Model x 182, reduced one-half. 6 Drawing of a reconstruction showing the topographical relation of the superficial thymus to the thymus head. These two structures, in the specimen from which the reconstruction of the thymus was made, are connected with each other by the connecting band (P.7.) which loops over the hypoglossal nerve. Left thymus, lateral aspect. 30, reduced one-half. 7 Drawing of model represented in figure 6; ventral aspect. 336 DEVELOPMENT OF THE THYMUS PLATE 2 J. A. BADERTSCHER 337 MITOCHONDRIA (AND OTHER CYTOPLASMIC STRUCTURES) IN TISSUE CULTURES MARGARET REED LEWIS AND.WARREN HARMON LEWIS! From the Anatomical Laboratory, Johns Hopkins Medical School, and the Marine Biological Laboratory, Woods Hole, Mass. TWENTY-SIX FIGURES CONTENTS IEAM PROKOUDIOTANC Nat... ane ght Stchits Gren artnet ig Mika scale Preuss Gisin 0 Ace eee hee bh: 340 WOGIMING, t.3 ore eRe acto eae Oh SOG ee ON ote feb rSlre: SOMO Ry a RE Dy oe Oe 341 RUDE ANOVA os 25 crete ate Ot Re SELEY oct ES A oy Me 344 WiitOochondriacese se athe otter coe ree eee sc ie Oe Del ae oe 347 JEONG Ol MMNUTCENOMNOITA, -cooocodonobooonsedodcietovndodausnoedarasonsucss 348 SHApeyOremitochOnaii Aas cance icy ce cde ene sy okete oA nea hn 352 STAC LOteNGOCH ON GIN ee Mica AA yuan anes cle eM es heat ee ee oe 359 NiumbersotemnbochoOndritiscpayc ests acticin noe ee os on ees 360 Quantity of mitochondria... 2st ash > ieee aot: Sake etn ce Doth ewes 360 Relation between position, size, number and quantity of mitochondria... 362 DECE MERLE mT OCMOMORA wrcha.cchea So Scene mul a Ae cota lrs henna hi aaa 363 INieoelacrielBiet Tah MANOS, dap ooooas eames & eceroeuie bide ase aoe Salminen one 365 Miro chomdmicnnmnrchiter ete kaimd Skoicelll Saas een nn ene 372 | Bisa graven ean TEM SHOT eR Ase A orp tage ne MCR Ra Ai Cet Ae Ree are eRe erm eels 373 TRIBE EL AKOTA) EXCISE ane Dee ae ae OER RS Re Sikes, SR eke nL Se bat ee eta 373 VCACuION SG Ona calTeSt eae ric ctaphin Orsitc ciongn 5 Aree ae ae etre, ot me aaa 374 Readchonsronaylolarchlorotormrether 4 ssee see eee ee ee ae eee 374 Reaction to hyper andthypotonic solutions:...J5.0. 0.5602 o> eens ene. 374 TEM SEY CILANGTEUHUG)LOVELEIES cael Mtl 5 elles aera ne 8 neg a ke Bab 374 WUTC AGREES 2 & ORM ey Ave 1 each ANA CU Ne Sg Me a 9 a OS 376 AIGHOV TIS) ARSE OF 5 NSC RNS. i sai CENSOR OP Ren MN PNENE ig ee Reco ete Ne ined Cm a ARETE 376 Nile blue B extra and brilliant cresyl blue 2b... 2.2.0 -...0.0..-25 0560s se: 377 GVO SVE degaene AED EA LS). 3's hn ON he eo a Oe Ey IRS, ey a 379 Certain other cell structures and their relation to the mitochondria........ 380 CGAL Sie en meee rey 0 We Wo nh OE URN Cae Oi aS kun Wa 380 NSC LOL CS heyaerce encom ste he pee MA hea A Ce yrs POR As ees 381 Retr clab lester enn sida We eh ak eee bale e Nawete ad Maly 386 Wait ea lecuele tar seins lee oe etl Aas a se in oes ce SE ens s See ree noe S os 390 ANUAOLOIST ES} CEU aKOUTENG 5 71 (G2) KC Ne Cray Ge Re a a 391 IDS ODISETVOIN SO Bis oS Sedan ie Say RE SY ys RO Rel ac 392 Conc lusionts sae pene ty ee nen Se oa gia aehi os Ase e304 EDO COED Lye TS EU NS 6G thik ie, ocsssle. sib os Aedes REG Oe ORR 397 1 We are indebted to the Marine Biological Laboratory for the use of a room during the summer of 1914. 339 340 MARGARET R. LEWIS AND WARREN H. LEWIS INTRODUCTION Tissue cultures afford a new and somewhat different method from that usually employed for the study of many cell struc- tures. It enables one to compare the living with the fixed mate- rial. In fact, one can study the same cell while living, during the process of fixation, and later as a stained permanent prep- aration. It also enables one to follow the changes which take place in the living cell from minute to minute. Above all, tissue cultures afford a method by which we can experiment on the cells and mitochondria. And through such methods only do we believe a correct interpretation of the significance of mito- chondria is to be found. In spite of the new and different environment of the tissue, i.e., Its isolation from the rest of the embryo; the substitution of a simple Locke’s solution for normal plasma; the contact with the cover-slip; and the absence of a circulation, which continu- ally renews the food-supply and removes the waste, the cells of the tissue cultures are apparently quite normal during the first two or three days and exhibit no noticeable changes except the characteristic configuration of the growth. How greatly the new environment disturbs the normal metabolic processes of the cell is impossible to surmise. The cells are in such a thin layer that each cell is probably as well bathed by the Locke’s solu- tion as in the embryo it would have been bathed by plasma or lymph. In the older cultures the cells lose their normal appearance and show signs of degeneration. Migration, growth and mitosis cease, the cells become smaller and show both cytoplasmic and nuclear changes. This may be due to the fact that the medium lacks both the inorganic and organic substances necessary for the prolonged continuance of life, but when we consider that the same degeneration takes place when tissues are explanted into a plasma medium it seems more probable that the degener- ation is due to an excess of waste products accumulated around the cell. MITOCHONDRIA IN TISSUE CULTURES 341 It is during the first two or three days then that we may compare the cells and their structures with those found in the embryo. The mitochondria have been studied during this early period when ‘their appearance and behavior can be considered normal. The close resemblance of the mitochondria found dur- ing this early period to those found in the chick by other ob- servers (Benda, Meves, Duesberg, Dubrueil, Cowdry, ete.) shows that they at least are not noticeably altered in the culture. We are justified, we believe in assuming that our findings concern- ing mitochondria apply as well to the normal cells within the embryo as they do to the cells of the tissue culture. TECHNIC The ordinary technic for the cultivation of tissues in Locke’s solution as described by Lewis and Lewis (’11, 12) was used. We found great variations in the amount, duration and character of growth in different solutions. This was apparently not due to the slight variation that occurs in the weighing out of the salts or sugar, which enter into the composition of the solution, since these can be varied considerably and good growth obtained. The trouble lies either in the distilled water, a contaminated container for the solution, poor chick material, or some manip- ulation during the process of explantation, which we vary un- knowingly. In repeating this work one should make several trials until a solution favorable for growth is obtained. When a favorable solution is once obtained it can be kept for months, provided the dextrose is not added’ to the stock solution. Chick embryos were taken out of the egg under aseptic condi- tions and put into 10 or 20 ce. of sterile Locke’s solution (NaCl 0.9 per cent, CaCl,0.025 per cent, KCl 0.042 per cent, NaHCO; 0.02 per cent, dextrose 0.25 per cent at 39°C. A piece a few milli- meters in diameter of the desired tissue was then cut out and placed in another dish which contained 10 or 20 ce. of sterile Locke’s solution at 39°C. This small piece was then cut up into numerous very small pieces. These were drawn up into 342 MARGARET R. LEWIS AND WARREN H. LEWIS a fine pipette, usually one at a time, with some of the solution and placed each on a sterile cover-slip which was inverted onto a vaseline (melting point, 46 + °C.) ring on a hollow ground slide. All instruments and cover-slips were sterilized by passing them through the flame, and aseptic precautions were observed throughout. Great care should be taken to insure absolute cleanliness of the cover-slips. The migrating and dividing cells, as we have already stated, adhere to the cover-slip and utilize it as a means of support, and the presence of grease seems to prevent them from getting a foothold. The small drop should spread out evenly and thinly over the center of the cover-slip so that the surface tension keeps the explanted piece in contact with the cover-slip. The stereotropic cells can thus easily creep out from the piece to the cover-slip on which they migrate towards the periphery of the drop. In cases where the drop is too deep and the small explanted piece falls away from the cover-slip the con- vex surface of the drop may act as a support for growth. Growth began within ten to twenty hours and reached a maximum in extent and showed the greatest number of mitotic figures on the second or third day. The cultures were incu- bated at 39°C. to 40°C. in an electric incubator (with a glass in the door). The presence of the electric light which was placed in the same chamber with the cultures for the purpose of main- taining the temperature of the incubator did not seem to affect the growth. Cultures apparently grow as well in the light as in the dark. Around the piece of explanted tissue the new growth forms a more or less radiating reticulum, a syncytium, or a membrane- like sheet of cells with varying numbers of isolated cells. The growth may be several cells in thickness near the old piece, but toward the periphery there is usually only a single layer of flattened cells which are often scarcely 2 u in thickness (fig. 1). The entire contents of these peripheral cells can be observed with very little change in focus. The growth is so closely at- tached to the cover-slip that in many cases the explanted piece can be torn away without injury to the new growth. MITOCHONDRIA IN TISSUE CULTURES 343 a. Pea a. Ay ~ 2 bs “S Yo me, > a,“ ' 6 iV EX : ‘ be ay = : "oo 2 ~ RD heh oL ye N, < ; : Vantes? eb wy he x Ch a >: “- f * Py : . uf —_— , BP eae Sak y Ode ae * « ‘ 2 Tr ane —~—* oe id Led - ow My hla )* Aloe JE * me *» ss Fe PI a a 9 : pas Selng ’ ar ~~ = A he : are ; h . * % N ~ Fig. 1 Photograph of part of a 2-day culture of intestine from an 8-day chick. The black mass is the explanted piece, which is surrounded by the new growth of connective tissue and smooth muscle; there are 13 mitotic figures in this part of the culture; osmic vapor and iron hematoxylin. Although several different kinds of cells have been identified in the living cultures (Lewis and Lewis) as, for instance, the mesenchyme and connective tissue cells, the heart and smooth muscle syncytium, the endodermal membrane, the yolk mem- brane, the nerve cell, the kidney tubule cell, ete., nevertheless, the general cytoplasmic structure of the living cell, regardless of the kind of cell, is practically the same except in cases where the cells contain secretory granules. The cytoplasm appears as a homogenous substance within which are several types of granules, i.e., refractive fat globules, various shaped mitochon- 344 MARGARET R. LEWIS AND WARREN H. LEWIS dria and other granules. The nucleus appears as a finely gran- ular body surrounded by a definite nuclear wall with one or more nucleoli. The nucleolus is never a round compact body, but instead is a coarsely granular ragged body, often large in proportion to the size of the nucleus. The nucleolus can read- ily be seen with the low power even when the outline of the nucleus cannot be distinguished. At one side of the nucleus there is usually present the central body (idiozome). When permanent preparations were desired the cover-slip was removed from the vaseline ring and the entire culture fixed to the cover-slip by means of osmic acid vapor. After fixation the explanted piece was often torn off from the cover-slip, leaving the new growth, in order to facilitate certain staining processes. Since the growth is very thin it was unnecessary to cut sections. The cover-slip with the fixed growth was treated as one would sections on a slide. ; FIXATION The entire process of fixation can be watched and studied upon any cell, as, for example, one that has been under obser- vation for some time. While the preparation is observed under the microscope some of the fixing solution can be introduced into the cavity of the shde through an opening made in the vaseline ring, which seals the coverslip to the slide. The speci- men can be fixed with either vapor or fluid. If a vapor is used, as of osmic acid, a small drop of a 2 per cent solution of osmic acid is introduced on the bottom of the cavity so that it does not come in contact with the hanging drop. If a solution is used, enough is introduced to fill the entire cavity and mix with the hanging drop. The vapor from a freshly made 2 per cent osmic acid solu- tion gave the best results, and when used with care resulted in a fixed cell, which more closely resembled the living cell than any other method we have used. The osmic acid vapour seems to cause a precipitation of the cell structures in the form of minute granules. Even after such fixed specimens are stained by means of iron hematoxylin the general character of the cyto- MITOCHONDRIA IN TISSUE CULTURES 345 plasm and nucleus are not noticeably altered except for the staining. B. F. Kingsbury (12) states that according to Rawitz (’07), Kollarewsky (’87) and Eisen (’00) osmic acid does not preserve nuclear details. So far as can be seen from our material the living cell exhibits few nuclear details and even osmic acid vapor differentiates more clearly the nuclear structures than they can be distinguished in the living cell. The mitochondria are so well fixed by osmic acid vapor or by a fixing solution which contains osmic acid that it has been suggested that the mitochondria may be artifacts due to osmic fixation. Vapor from strong formalin which has been carefully neutralized (Mann ’02 and also Bensley °11 recommend that formalin be freed from acid by careful neutralization and redis- tillation) gave good results in regard to fixation not only of the mitochondria but of all cellular structures. Unfortunately, the mitochondria did not stain well after formalin fixation. Iodine vapor from a crystal of iodine often afforded good results in regard to the spindle fibers and also the mitochondria, especially so when followed by Bensley’s anilin fuchsin, methylene green stain, but 1odine was an uncertain fixative. Osmie acid solu- tions do not give as uniformly good results as the vapor. Any fixing solution which contained acid (acetic, hydrochloric, sulphuric, etc.) proved useless as a fixative for tissue cultures. The vapor from such acids coagulated the entire cell before the fluid touched the preparations. The mitochondria rapidly changed into small granular rings, which later were completely lost in the coagulated network of cytoplasm. The nucleus lost its homogenous finely granular structure and a coarse network appeared. The nucleolus became a small round body. This re- sembles closely the usual textbook figure of a cell, which by long association one has come to believe represents a cell but which actually resembles the living cell not at all. When a living cell (fig. 2a) was exposed to the action of vapor from 2 per cent glacial acetic this coagulation effect was soon apparent, as shown in figure 2b. The cytoplasmic and nuclear networks rapidly appeared while the mitochondria, which 346 MARGARET R. LEWIS AND WARREN H. LEWIS >= —sS \ = loavele acme Oui ee 11.02r 11.03 11.04 am 12.54pm 12.55 pm 12.56 pm e TTL LLY fo FF 1.30 pm 1.34pm 1.36 pm L 10 Het ate a eile ee ate E15) elon eG ael6r pm b d [eo soa on oe Se See See ee C- ca a 11.16am 11.17 11.18 11.19 11.20 11.21 11.22 11.23 am Cc 11 pm nee a bs Vas : 1.8 pm Fig. 2. A, sketch of living cell with nucleus and mitochondria, the cytoplasm should be homogenous. b, the same cell after exposure to the vapor of 2 per cent acetic acid; both cytoplasm and nucleus show reticular structure due to coagulation; the mitochondria are destroyed except for a few granular remains. Fig.3 Changes in shape of mitochondria as observed in living cells; a, changes by bending during a period of 4 minutes; b, by shortening and elongating and shifting along the mitochondrium of the mitochondrial substance, time 6 minutes; c, by fusion of granules; in the course of 7 minutes the four granules a, 6, c, d became fused into two granules ab and cd; a, b, c from a 2-day culture of mesen- chyme from a 43-day chick; d, changes in shape and fusion of mitochondria to form network in the course of 6 minutes; e, changes in shape and fusion, 2 minutes; /, changes in form of a single mitochondrium during a period of 7 minutes; d, e and f from a 3-day culture of mesenchyme from a 4-day chick. MITOCHONDRIA IN TISSUE CULTURES 347 were clearly seen in the living cell as long rods and threads rapidly disappeared. Brunn (’84) found that the Eberth bodies, which have since been shown to be mitochondria, are dissolved by acetic acid. Duesberg (’11) states the fact that they are destroyed by acetic acid to be one of the criteria for mitochon- dria. Prolonged fixation in osmic acid after the osmic vapor has already fixed the preparation did not cause any distortion of the nucleus or the cytoplasmic structures. The mitochondria became somewhat blackened and the fat globules were first yellow-brown and later became a dark brown. Even after a month no change appeared in the cytoplasm which in any way indicated the presence of a canilicular apparatus as was found in certain other cells by Kopsch (’02), Sjovall (06), and Cowdry (712). A careful study of the living cell together with a study of the effect of various fixatives shows that while the mitochondria are only successfully fixed by osmic acid, they are by no means artifacts due to osmie acid fixation. MITOCHONDRIA The mitochondria are always present in the living cells of the tissue cultures and after some experience can be easily recog- nized. They are never as conspicuous, however, as the fat globules or the nucleus. The mitochondria are slightly refract- ive bodies which vary greatly in shape, size, and position. In the living cell these bodies are never quiet, but are contin- ually changing in shape, size, and position. Often as many as fifteen or twenty shapes may be exhibited by a single mitochon- drium within as many minutes (fig. 3). This extreme plasticity of the mitochondria is a very important characteristic and was shown in every preparation examined. It is certainly a feature which must be reckoned with in any attempt to classify or to analyze their behavior from fixed material. The chaotic condition of the literature in respect to the ter- minology and criteria for mitochondria and other cytoplasmic bodies renders it difficult often to correlate our findings with 348 MARGARET R. LEWIS AND WARREN H. LEWIS those of other observers. It is important then that we should as far as possible submit the bodies herein under consideration to already established criteria for mitochondria. While these bodies fulfil Benda’s (’99) original criterion for mitochondria in embryonic cells in that they stain blue with alizarine, we have made no effort to fulfil Montgomery’s (11) criterion that they must show an unbroken cycle from egg to somatic cell to anlage sex cell and back to the fertilized egg. They do, however, correspond with Duesberg’s (711) criterion for mitochondria in the adult cell, in that they are seen in the fresh preparation, dissolved by acetic acid, preserved by osmic acid, and stain by the same dyes as the mitochondria in em- bryonic cells, that is, green with Janus green in the fresh prep- arations (Michaelis ’99, Laguesse ’99, Bensley ’11, Cowdry ’12); stain blue with Benda’s stain (Benda ’03, Meves, ’08, Duesberg 09); red with Bensley’s anilin fuchsin, methylen green stain (Bensley ’11, Cowdry 712), and black with Heidenhain’s iron hematoxylin. Janus green caused the death of the growth after a few hours, and frequently the mitochondria separated into granules (fig. 21). For this reason Janus green was used only to identify various granules as mitochondria but never for any observations upon the changes in shape, size or quantity of mitochondria. These bodies have been given various names—mitochondria and chondriomiten by Benda; chondrioconten, chondriosomen, chondrion and plastosomen by Meves; plasmafaden, plasma- k6ren by Retzius; paramiton or miton by Flemming; micro- somen by Van Beneden; granules and filament by Altman, ete. Position of mitochondria Great variation occurs in the arrangement of the mitochondria even in the same kind of cells in the same preparation, not only in the living but also in the fixed preparations. It is not uncommon for the mitochondria to be more or less evenly scattered throughout the cytoplasm and the various processes of the cell. They have been observed even in the extremely MITOCHONDRIA IN TISSUE CULTURES 349 slender processes that are scarcely larger in diameter than a mitochondrium. This rather uniformly scattered arrangement usually occurs during mitosis and here likewise the cell processes may contain mitochondria. The spindle area is usually free from mitochondria (figs. 4, 15, 16, 17). Infrequently in the late anaphase the mitochondria may collect along the plane of division. In elongated cells the mitochondria are usually arranged at either end of the nucleus with their long axis more or less parallel to the long axis of the cell. However, in many of the cells the mitochondria are more numerous about the nucleus or about the central body than towards the periphery of the cell, where they may be scattered or entirely absent (figs. 4, 5, 12, 17). The central body is an extremely finely granular body at one side of the nucleus and has been so-called by us because the mitochondria frequently radiate around this body and because it is a non-committal term. Usually the idio- zome (or nebenkern, for discussion of correct terminology see Wilson 711) can be seen within this body and occasionally the centrasome can be made out within the idiozome. The central body is more clearly seen in the living cell than in the fixed cell, but in some cells this body cannot be distinguished and the mitochondria appear more or less radially arranged around the nucleus. At times the mitochondria may be confined to one side only of the nucleus, usually the side on which lies the cen- tral body. This radial arrangement about the central body has been described by Eberth (66), Vejdovsky (07), Meves (09), Veratti (09). In some preparations this arrangement is so marked that one cannot but wonder if there is not some definite relationship between the two, and it is not difficult to under- stand why Vejdovsky (’07) believed that the mitochondria were products of the activity of the centrasome. The mitochondria, however, are continually altering their posi- tion, not only in relation to the nucleus and central body but also in relation to each other. They seem to be continually emerging from the mass near the nucleus or near the central body and to migrate out towards the periphery. Also those towards the periphery often return to the central mass. There THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No.3 350 MARGARET R. LEWIS AND WARREN H. LEWIS \¢ LING 3 Giees wv JOSME Noe 2 /-: Os ee er se SEA GGL ee, cay \’ Oe OOS AS AS Ae fall hee mat aca Wy f Oe SL fo € ae ai a \ Sh : BNNs / = ie \ \ \ ili 1 aS » \ / \ Fig. 4 A, b,c, d,e, cells from a 2-day culture of heart from a 5-day chick. a, cell with 69 mitochondria of granular type, somewhat radially arranged about the central body. 6, cell with 125 mitochondria of very different shapes and sizes, about the central body and scattered through the cytoplasm. c, cell with 37 mitochondria about the central body, mostly granular in type; X 1080 diam. d, cell with about 90 mitochondrig about nucleus and central body, mostly rod- and thread-shaped; X 540 diam. e, group of six adjoining cells; as in a the cells with granular and short, rod-shaped mitochondria show the latter arranged about the nucleus and central body; one cell contains 40 and the other 54 mitochondria of the short and long rod- and thread-shaped types, which are arranged more MITOCHONDRIA IN TISSUE CULTURES 351 about the nucleus and central body; the dividing cell has 118 mitochondria which are scattered more evenly through the cytoplasm than in the other cells; X 790 diam.; osmic acid vapor and iron hematoxylin; f, cell from a 2-day culture of heart from a 6-day chick cell with 152 mitochondria, mostly rod- and thread- shaped, arranged about nucleus and central body; X 790 diam. g, cell froma 2-day culture of heart from a 4-day chick. Only two cells in entire culture show these ring-formed mitochondria and in these two the cytoplasm was abnormal, perhaps dead. Bensley stain; X 790 diam. Fig. 5 A, 6, c,d, four adjoining cells from a 2-day culture of heart from a 5-day chick; X 790 diam. The resting cells, a, b and c, have 47, 51 and 48 mito- chondria respectively, while cell d in early prophase has 89. These cells exhibit great variety in the shapes of the mitochondria; small and large granules, spin- dles, short rods, long rods and threads are present; the mitochondria are some- what more scattered than the cytoplasm in cell d. e, cell with 38 mitochondria which vary greatly in size and shape; from the same specimen; X 540; osmic acid vapor and iron hematoxylin. /, g, mesenchyme cells from 2-day cultures of intestine from a 7-day chick embryo; f, many of the mitochondria are united in networks; < 540 diam.; g, some ot the mitochondria branch and anastomose to form a complicated network which appears to extend from one nuclear area to another, where the cells form a syncytium; X 790 diam.; only a portion of the cells and network is shown; osmic acid vapor and iron hematoxylin. 352 MARGARET R. LEWIS AND WARREN H. LEWIS is probably no relation between the movement of the cell and the movement of the mitochondria, for the cell processes change their position so slowly that there is often no noticeable change for several hours, while the mitochondria change position rapidly and continually. Occasionally in a syncytium of cells a mitochondrium may pass over the cytoplasmic bridge from one cell to another (fig. 5g). In some cases a mitochondrial thread may pass over the bridge into another cell and later return. What is it that governs the arrangement of the mitochondria? Is it the shape of the cell, the influence of the central body or “of the nucleus, the internal structure of the cytoplasm, or do the metabolic activities of the cell govern the size, shape and arrangement of the mitochondria? Shape of mitochondria The mitochondria exhibit extraordinary diversity of form often in the same preparation, even in adjoining cells of the same type (figs. 4, 5). Not infrequently a single cell may contain mitochondria of diverse shapes (fig. 5e). These different mito- chondrial shapes may be more or less localized in different parts of the cytoplasm (fig. 5 e) or may be more or less mixed together (fig. 4b). The extraordinary diversity in form of the mito- chondria shown by cells of the same type lying side-by-side in the same preparation is sometimes very striking. Such differ- ences occur in the young growing cells after division, in older resting cells and even during the various stages of mitosis. Again, we may find in the same preparation groups of cells in one part of the growth, that have very similarly shaped mitochondria, while in another part practically all of the cells may have quite differently shaped mitochondria. In such prep- arations all gradations in shape and size, from minute granules to larger and larger ones, or from rods to threads of various lengths, or threads and networks, etc., can be seen in adjoining cells of the same type or even in the same cell. Just as the fixed preparations show such gradations, we find that all sorts MITOCHONDRIA IN TISSUE CULTURES abs of transformations from one shape into another can be watched in the living cell. Mitochondria of various shapes have been described by other observers, and so definite did some of the shapes appear to be that they were given various names, which today are without much significance. Nevertheless, it is convenient to classify mitochondria as follows (fig. 10). Small granules Threads Dumb-bell-shaped granules Loops Spindle-shaped granules Rings Large granules Network Rods The degenerate mitochondria also show more or less definite shapes (fig. 13). Mitochondria continually change shape as by bending in vari- ous directions (fig. 3a), or by shortening and thickening or elongating and thinning (fig. 3b); at times this thickening and thinning seems almost like a pulsation along the length of the mitochondrium. These various shapes of mitochondria are not fixed or constant in any cell. Rods or threads may change into - granules; threads may fuse or branch into networks (fig. 3 d, 6, 7); or granules may fuse to form larger granules (fig. 3c). Degen- erating mitochondria may separate into granules and _ vesicles (fig. 13). Ring-shaped mitochondria are seldom found in these prepa- rations. Occasionally a living cell may contain one or two large or small ring-shaped mitochondria which rapidly change into threads, rods or granules. A few fixed and stained preparations show one or two cells at the periphery of a large growth which contain ring-shaped mitochondria exclusively (fig. 4g). Kings- bury (11) has suggested the possibility that mitochondria which contain a large amount of lipoid are reduced by osmic acid only at the surface, and the central part later dissolves out, which produces the appearance of rings. These ring-shaped mito- chondria can hardly produce fat or lipoid droplets (Dubreuil 11, 718) since they are seldom present in cells in which fat is being formed. 354 MARGARET R. LEWIS AND WARREN H. LEWIS a Fe ee ee ~~ 4.00 P.M Fig. 6 From a living culture about 24 hours old, of a piece of heart from an 6-day chick, showing branching, fusion and splitting of two or three mitochondria during a period of about 1 hour. MITOCHONDRIA IN TISSUE CULTURES 355 ae \: “ ate a Ne z + © \o a Soe a t b 2.40 P.M 2.42 P.M. 2.43 P.M 2.45 P.M. .S Lo . ok: AIS Ke Y * t 0. 2 2.46 P.M. 2.47 P.M. 2.48 P.M. 250PM. Hi EN 5 A o )* WV D> al © : (oz per eas ZEN / Jes ei 7 4 Ua a Se g nee 2.55 P.M. 2.56 P.M. 2.57 P.M. 2.58 P.M. 3.00 P.M 3.02 P.M. 3.05 P.M Fig. 7 Changes in shape and anastomoses of a few mitochondria during a period of 25 minutes in a living mesenchyme cell. The changes were so rapid that it was not always possible to draw each mitochondrium; they can be fol- lowed by the lettering; 24-hour old culture from a 6-day chick embryo. The mitochondria are frequently arranged in the form of a network (fig. 5 f, g) which may involve many of the mitochondria or only a few of them. A study of the fixed preparations and especially a study of the living cells shows conclusively that Mislavsky (11) is correct in his contention that the mitochondria do fuse and branch to form networks. We have observed all 356 MARGARET R. LEWIS AND WARREN H. LEWIS stages of the formation of network in the living cell. These networks continually change shape (figs. 6, 7). New branches appear, old ones change shape or position or break away, and at times the entire network may break down into loops, threads and granules without any apparent change on the part of the cell. From our observations it appears that the network is Fig. 8 Mitochondria from a 3-day culture of intestine from a 7-day chick; osmic vapor and iron hematoxylin; X 2250 diam.; various forms of mitochondria, which come from the breaking down of a network, into loops, rings, threads, etc. very unstable and rapidly breaks down into granules, loops and threads. Figure 8 shows such loops and rings in a fixed specimen. There has been some discussion as to which shape of mito- chondria is the more primitive. Meves (’08) claimed that in the twenty-hour chick embryo the mitochondria are present MITOCHONDRIA IN TISSUE CULTURES 307 only as very thin threads, but that at forty-eight hours the threads are thicker and also some granules are present. These mitochondria have heavy stained edges with a clear mark substance. Duesberg (’08) finds the same for the chick, but in the rabbit he describes the mitochondria in the early fertilized egg as small granules, which increase in volume and become large granules at the end of the third day. The large granules have a clear central part and dark outer edge. They flow together and build rods and threads. Rubaschkin (’11) finds only granules in the early guinea-pig development. He claims that the granular form of mitochondria is the most primitive and indifferent form. So far as could be observed, there is no special difference in the shape of the mitochondria present in the cells of the growth from a piece of a three-day chick embryo from that present in the growth from a piece of ten-day chick embryo. Only those cells show exclusively the small granules, which contain many fat globules or vacuoles. We have observed the cells of a 51 hours growth which contained only the granular type of mito- chondria to contain at 70 hours mostly thread types (fig. 9 a-f). The threads were formed by the stretching out of the granules rather than by fusion of granules although such fusion of granules does take place. When a preparation is studied from day to day it is clear that the shape of the mitochondria changes and that no one shape is constant for any one age. Brown (’13) finds that in the male germ cells of Notonerta the mitochondrial fibers and threads arise in part at least from spherical-shaped mitochondria. : Schaxel (’11) claims that the shape of the mitochondria varies with the method of fixing and staining inasmuch as by the Benda treatment the rod-like forms predominate while after the Altman treatment the granular type predominates. While there have been few observations made as to the effect of various technical methods upon the shape of the mitochondria they appear to be such malleable structures that it is quite probable that their shape could be altered by different methods. 358 MARGARET R. LEWIS AND WARREN H. LEWIS Pi g 4.50. P.M. h 4.55 P.M i 500P.M j 5:04PM kK 5.08 P.M {5.15PM Fig. 9 Mitochondria in living cell from a culture of heart from a 43-day chick. a, camera drawing at 2.20 p.m., when culture was 51 hours old; b, same cell at 3.20 P.M.; c, at 4.20 p.m.; d, at 4.55 p.M.; e, at 5.15 p.m.; at this time all the cells had a similar type of mitochondria; f, camera drawing at 9.20 a.m. of next day, culture 70 hours old. All the cells in the culture had the same type of mito- chondria asine. At4 p.m. most of the cells began to show signs of degeneration and a fresh drop of Locke’s solution was put on preparation; the thread-like mitochondria, as seen in f, began to fuse into large spindle-shaped masses near the nucleus and central body, as in g, where all the mitochondria now present at 4.50 p.m. are shown, and the changes which they underwent during the next 25 minutes are in this particular cell shown in h, 7, j, k and l. MITOCHONDRIA IN TISSUE CULTURES 359 Size of mitochondria The mitochondria vary so greatly in size (fig. 10) that were it not for prolonged study of them and the use of a specific vital stain such as Janus green it would be difficult to believe that they all belong in the same class of granules. Even in a single cell great variation occurs from very minute granules which are scarcely visible to relatively large masses (figs. 4, 5, 9). IO ONL | {Le roniters Fig. 10 Camera lucida drawings of-mitochondria of various sizes and shapes from different cells and specimens; osmic acid vapor and iron hematoxylin; X 790 diam. Fig. 11 Endodermal cells from 3-day culture of allantois from a 7-day chick; Bensley’s aniline fuschsin methylene green stain; X 790 diam. Occasionally a cell is seen in which all the mitochondria appear to be swollen up and much larger than those in the surrounding cells (fig. 11). To what this is due is not known. A mitochondrium under observation frequently seems to change in size as well as shape, but so far no micrometer measurements have been made to determine this point. Definite increase in size has frequently been seen, due to fusion of two granules to form a larger granule or to fusion of rods into threads; and occasionally all the mitochondria in the cell may become col- 360 MARGARET R. LEWIS AND WARREN H. LEWIS lected in several very large granules (fig. 9). In certain patho- logical conditions Barratt (13) has found that the mitochondria become abnormally large and stain clearly. Number of mitochondria The number of mitochondria varies greatly in cells of the same kind in the same preparation (fig. 12) and in different preparations (figs. 4, 5, 12). Numerous counts of the mito- chondria in the same kind of cells in the same preparation show that there is no one number of mitochondria peculiar to any one kind of cell or to any one stage in the development of the cell. The number of the mitochondria appear to decrease and to increase under various conditions. This may result from fusion or division of the mitochondria without much change in the quantity of mitochondrial substance; or this may be accom- panied by a corresponding increase or decrease in the amount of mitochondrial substance, independent of any fusion or divi- sion of the already existing mitochondria. This would indicate that some of the mitochondria may at times entirely disappear and that possibly new ones may arise de novo in the cytoplasm. Sometimes most of the cells in a growth undergo such changes. When observed on one day they may have rather few mito- chondria, while on the following day most of the cells may contain a marked increase in the number of mitochondria, or the opposite phenomenon may take place. This may or may not be accompanied by a corresponding change in the quantity of mitochondrial substance. Prolonged action of heat causes a decrease in the size and number of the mitochondria, and it is hoped that further experimental work will determine what conditions cause such changes in the ordinary cultures. Quantity of mitochondria By the quantity of mitochondria we mean the total mass of the mitochondrial substance within a cell. This can only be roughly estimated, as some cells with many very small mito- chondria have a smaller quantity of mitochondrial substance than others with fewer but larger mitochondria. However, in MITOCHONDRIA IN TISSUE CULTURES 361 Fig. 12 A, mesenchyme cells from a 2-day culture of intestine from a 5-day chick, showing marked differences in shape, size and number of mitochondria. The four cells have 74, 8, 27 and 6 mitochondria; Bensley stain; X 790 diam,; b, two adjoining cells from a 2-day culture of heart from a 7-day chick; granular type of mitochondria, one cell has 38 and the other 111 mitochondria. Osmic acid and iron hematoxylin; < 540 diam.; c, two endodermal cells from a 2-day culture of allantois from a 4-day chick; the larger cell contains about 128 and the other 27 mitochondria; Bensley stain; x 790 diam. many cases of adjoining cells (figs. 11,12b) or of the same kind of cells in different parts of the preparation (fig. 12 ¢) there can be no doubt that the quantity of mitochondria is markedly increased or very much decreased. This increase in the quantity of mitochondria is most marked in a few scat- tered cells in the growth from a piece of allantois (fig. 11). Cells with few mitochondria do not necessarily have larger mitochondria and there seems to be no definite relation between size and number or number and quantity. The quantity in the cell differs so widely that it has so far been impossible to con- nect the quantity of mitochondria with any one factor. Possibly it is dependent upon the metabolism of the individual cell. 362 MARGARET R. LEWIS AND WARREN H. LEWIS This is also true of cells undergoing division, for there seems to be no amount of mitochondria characteristic of any one phase of division. Also the variation in the quantity of mito- chondria present in any one phase of division is considerable, as can be seen (figs. 14, 15, 16). Daughter cells usually have a smaller quantity of mitochondria than the metaphase cell or than the resting cell (fig. 17). As to the question whether the amount of mitochondria in- creases during mitosis it is impossible to state. So far, we have only one definite observation that this is true. In this case the living culture was subjected to a temperature of 46°C. for two hours, during which period the mitochondria decreased decidedly in number and size. Two cells which were under observation suddenly began to pass into prophase and during this process the number of mitochondria in these two cells increased until they contained more than they had before the experiment was begun. Although several subsequent experiments with increased heat caused a decrease in the quantity of mitochondria no cell division was observed. No agent but heat has so far been observed which caused a change in the amount of the mitochondria without injury to the cell. However, it is evident that certain metabolic condi- tions must cause a change in the quantity of mitochondria. Relation between position, size, number and quantity of mitochondria No definite relation between the position, size, number and quantity of the mitochondria has been observed in the cells of the tissue cultures, still there is a more or less marked manner in which the mitochondria occur in the cells. Frequently the long threads or short rods are plentiful and scattered throughout the cytoplasm with or without a definite central body. When the mitochondria are in the form of large granules and thick rods they are fewer in number and are arranged more or less radially around a central body. When only a very few mitochondrial granules are present they are usually of the large granule type. MITOCHONDRIA IN TISSUE CULTURES 363 All phases in the development of the cell, i.e., daughter cell, growing cell, resting cell and dividing cell can be found with any one of the above combinations of the mitochondria. How- ever, it must not be forgotten that many cells contain part of one kind of mitochondria and part of another and that any one shape of mitochondria may turn into another at any time during observation, and that no one shape of mitochondria remains as such for a very long interval of time, but changes into another. hd i ee 7 ————. — <> b aS \ N 138 a hae a te © hbe oes 345 Pm e 3247 em Fig. 13 A, cell from a 2-day culture of heart from a 7-day chick; practically all the mitochondria are degenerated, those in the region of the central body show most advanced stage of granular rings; X 790 diam.; b, degenerating mito- chondria from mesenchyme cell of a 2-day culture of intestine from a 4-day chick; X 920 diam.;c, cell from a 2-day culture of heart of a 5-day chick; all the mitochondria have degenerated into granular rings; osmic acid vapor and iron hematoxylin; X 790 diam.; d, process of degeneration in a single mitochondrium produced by the action of acetic acid vapor on a living cell; e, effect of CO, on another mitochondrium in 2 minutes. Degenerate mitochondria Degenerate mitochondria of various shapes are occasionally found in these preparations (fig. 13). A study of the cells of the older growths shows that all the mitochondria do not neces- ‘sarily degenerate at the same time. Some cells are found which contain many normal mitochondria, some partly degenerate, and others entirely degenerate. 364 MARGARET R. LEWIS AND WARREN H. LEWIS This degeneration appears first in the mitochondria around the central body and later in those scattered at the periphery (fig. 13a). The process of degeneration of the mitochondria can be most successfully observed when produced by some outside agency such as carbonic acid gas or vapor from a weak acid solution (fig. 18d, e). When the death of the cell is produced experi- mentally the mitochondria become first a series of granules which soon become slightly vesicular although at this stage they still stain in the characteristic manner. Then these vesicles sepa- rate and rapidly become small, finely granular rings or shadows. These no longer stain like mitochondria but more like the cyto- plasm, i.e., brownish green with Bensley’s anilin fuchsin, methylen green or pale gray with Heidenhain’s iron hemo- toxylin, and in the living cell the Janus green does not stain them green. It is apparent that some change has taken place which has completely changed not only the morphology but also the composition of the mitochondria. These degenerate mitochondria correspond in many ways to the ‘‘grains du segregation’? described by Dubreuil in the lymph cells, but are unquestionably degenerate mitochondria, and they can be produced in any cell of these growths by means of various agents such as carbonic acid gas, chloretone, acid vapor, hydro- gen peroxide and potassium permanganate. Meves (710) and Duesberg (710) simultaneously found that poorly fixed mitochondria show granulation and small bladder forms. Other observers have found that granulation is due to delay in fixation after death or to disease, as Mayer and Rathery (07) experimental polyuria; Takaki (07) polyuria or prolonged fast; Policard (’10) experimental poluria and after injection of phlorizin; Policard and Garnier (’07), Cesa Bianchi (’10), Heiden- hain (711) also obtained similar results. Beckton (’10) claims that in a certain tumor no mitochondria were present in the tumor cells. In view of the observations of Beckwith (14) it may be possible that certain cells can exist without mitochondria, but it seems more probable that the apparent lack of mitochondria in the tumor cell described by MITOCHONDRIA IN TISSUE CULTURES 365 Beckton may have been caused by delay in fixing the material so that the mitochondria became degenerate, or the mitochondria may have been present only as degenerate structures which did not stain. Mitochondria in mitosis Naturally, the question at once arises: What is the rdéle of the mitochondria during division of the cell? Many observers be- lieve that the mitochondria form a palisade about the spindle during late anaphase and then divide and one-half of each mito- Fig. 14 Arrangement of mitochondria in the prophase stage; a, from a 3-day culture of intestine from an 8-day chick embryo; X 1080 diam.; b, from a 3-day culture of intestine from a 7-day chick; X 790 diam.; c, from a 2-day culture of heart from a 5-day chick embryo; X 540 diam. chondrium passes to each daughter cell (Benda, Duesberg, Meves, etc.). Meves (713) in his work on ascaris egg goes so far as to state that not only are the mitochondria present in the egg and spermatozoon, but also that the male mitochondria are carried into the egg by the spermatozoon and so each egg receives not only female but also male mitochondria and the granules result- ing from the fusion of the male and female mitochondria are distributed to each cell of the embryo. In view of the behavior of the mitochondria Meves suggests that they may play a part in inheritance. THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 3 366 MARGARET R. LEWIS AND WARREN H. LEWIS A study of the fixed specimens seems to show that the mito- chondria retain somewhat their original character and shape during mitosis (figs. 14-17). They are, however, almost always shorter and more scattered through the cytoplasm than in the surrounding cells (figs. 4, 5, 14 a, 15 b, d, 17, 18). There are usually as many and often more mitochondria in the early stages of the dividing cell than in the neighboring cells (figs. 4, 5, 14a, b, 15:b). There is no indication in the fixed specimens of any arrange- ment of the mitochondria about the spindle in such a manner that they would undergo division into two parts in the plane of cleavage of the dividing cell. On the other hand, all of our specimens seem to show that the mitochondria tend to become more evenly scattered through the cytoplasm during division, and those that happen to be on either side of the cleavage plane are carried into the respective daughter cells. Since each daughter cell contains only about one-half the number of mitochondria found in the mother cell at the time of division we must assume that there is an increase sometime during the life of the cell between one division and the next, otherwise the number would rapidly decrease during each suc- cessive division. Now the question is: When does this increase take place? Is it during the so-called resting period, or during mitosis? In some of the fixed preparations where mitotic figures, daughter cells and young growing cells are numerous, it is possi- ble to arrange cells in a series according to the stage of recon- Fig. 15 Arrangement of mitochondria during metaphase; a, b,-c, f, cells from 2-day cultures of heart from 5-day chick embryos; d, e, cells from a 3-day culture of intestine from an 8-day chick embryo; X 540 diam. Fig. 16 Arrangement of mitochondria during anaphase and telephase and young daughter cells, a, b, c, from a 3-day culture of intestine from an 8-day chick; X 540 diam. Cell a, anaphase has 156 mitochondria, the two daughter cells, b, have 12? and 125 each, while the older daughter cell, c, has 151 mito- chondria; the neighboring adult cells in this region have been 70 to 160 mito- chondria; d, 3-day culture of intestine from a 7-day chick; the two daughter cells with the smaller dark nuclei have 92 and 49 mitochondria, while the adjoining resting cells have only 77 and 48 mitochondria each; * 790; e, daughter cell from a 2-day culture of heart from a 5-day chick, with very different type of mitochon- dria; X 540 diam.; osmic acid vapor and iron hematoxylin. MITOCHONDRIA IN TISSUE CULTURES 367 368 MARGARET R. LEWIS AND WARREN H. LEWIS struction of the nucleus, as indicated in figures 17 and 18. The younger nuclei are smaller, darker and more compact and the cells are smaller. The older cells are larger and contain larger nuclei which are less and less deeply stained. In such a series (fig. 17) the number of mitochondria increases from about 40 to 150. On the other hand, the old resting cell (k) with a very pale nucleus has only 32. In one series (fig. 181, j, k, 1) the number increases from 24 to 140. In figure 18 one of the two daughter cells (ce) has 37, the dividing cell (a) has 140, while the two neighboring resting cells (e) and (f) have 39 and 47. Again, in figure 18, the young daughter cells (d, d) have 37 each while older neighboring cells (g, h) have 56 and 58 each. On the other hand, another dividing cell (b) near this same group has but 60 mitochondria. From such observations one might conclude that there is a gradual increase in the number and in the size of the mito- chondria during the growth period of the daughter cells. The greatest increase both in number and size seems to occur then during the so-called ‘resting’ period which is in reality a period of growth both for the mitochondria and for the nucleus. On the other hand, while we are unable to determine definitely whether the number of mitochondria actually increases during the early stage of mitosis there are frequently indications that such cells have more mitochondria than mature cells Cell d (fig. 5) early prophase has 89 while the resting cells a, b, ¢ have 47, 51 and 48 mitochondria each. In figure 4e the dividing cell has 118 while the three neighboring cells have 102, 126 and 62 mitochondria each. The two cells with the larger nuclei are probably older resting cells and each has about the same number of mitochondria as in the dividing cell. Numerous other specimens seem to show that the dividing cells often have more mitochondria than any of the fullgrown resting cells in the immediate neighborhood. Sometimes this is so marked that there is every indication that the number of the mitochondria in some instances may increase considerably dur- ing mitosis. It seems probable thereforee that mitochondria increase in number both during the resting period and during MITOCHONDRIA IN TISSUE CULTURES 369 Fig. 17 Cells from a 2-day culture of heart from a 5-day chick embryo; X 540 diam.; a, very young daughter cells with 42 mitochondria in each cell; 6, slightly older daughter cells with 42 and 35 mitochondria; c, older daughter cells with 48 and 75 mitochondria; d, older daughter cells with 43 and 38 mitochondria; e, older cells with 66 mitochondria; f, older cell with 96; g, still older cell with 156; h, prophase with 197 (?); 7, metaphase with 117; 7, anaphase with 174; k, old resting cell with only 32 mitochondria. 370 MARGARET R. LEWIS AND WARREN H. LEWIS mitosis; perhaps in some more during the resting period; in others more during mitosis and in still others during both periods or only during one. It is very unlikely that one can arrive at a satisfactory solution of such a problem from fixed material, since the bodies we are dealing with are subject to such great changes in number and size during life. The number of mito- chondria is not of much value as an indicator of the total quan- tity of mitochondrial substance. Fig. 18 Cells from a 2-day culture of heart from a 5-day chick; X 540 diam. Cell a in anaphase has about 140 mitochondria, while b has only about 60, the young daughter cell c has 37, the two daughter cells d, d have 36 each; the older cells e, f, g and h have 39, 47, 58 and 56 mitochondria respectively. In the series 1, j, k, l, the young cell 7 has only 24, the older cells 7 and & have 41 and 42, while the mature cell / has 140 mitochondria. We have indicated that the daughter cells not only have about one-half the number of mitochrondria found in the mature cells, but that the mitochondria are sometimes smaller. Does the increase in number during the growth period come about through division of preéxisting mitochondria (a process which frequently takes place) or do mitochondria arise de novo? So far as our observations go, either or both processes may occur. MITOCHONDRIA IN TISSUE CULTURES Sat The only certain method to determine just how and when the mitochondria increase is to follow several living cells through complete cycles. Unfortunately, the cells of tissue culture often round up during late metaphase and anaphase (fig. 15 e, f) so that it is impossible, except in a few cases, to follow the indi- vidual mitochondrium throughout cell division. The process of mitosis is an exceedingly slow one compared with that described in other tissues: Prophase 10 to 20 minutes, metaphase and anaphase, 1 to 2 hours; while the period from anaphase including telophase to the daughter cells is an exceed- ingly short one, never more than five minutes from the time the chromosomes are arranged at the opposite poles of the spindle until the cytoplasm is divided, except for slender processes, and such stages are correspondingly few in number in the permanent preparations. We have not been able to follow the number of mitochondria through a ‘complete cycle of the cell in the living cultures. We have, however, been able to watch the behavior of the mito- chondria during mitosis in a few living cells. Usually the mito- chondria are scattered throughout the cytoplasm and remain so during cell division. About one-half of the mitochondria pass to each daughter cell, namely, those which happen to be on one side or the other of the cleavage plane. In two or three cells during late anaphase most of the mitochondria became arranged in rather of a broad zone around the spindle in the area through which the division plane later formed and one-half of the number of mitochondria passed into each daughter cell. There was no indication of any division of the mitochondrial granules;'in fact, in one cell it was clearly observed that several thread-shaped mitochondria passed over entire into one of the daughter cells. A division of the mitochondria such as observed by Meves (’08) and Duesberg (’10) was never observed. We find as did Buch- ner (09, 710, ’11) that this characteristic arrangement of the mitochondria during division of the cell is by no means a con- stant occurrence. We have already stated that we are uncertain whether there is an actual or only an apparent increase in the amount of 372 MARGARET R. LEWIS AND WARREN H. LEWIS mitochondria during mitosis. So far we have only one direct experimental observation to offer, and in this particular case there was an actual increase in the number and possibly also in the quantity. In this experiment the temperature had been raised from 39 to 46°C. and was retained at 46°C. for two hours. There resulted a decided decrease in the amount of mitochondria within all the cells. Two cells began to divide. The nuclear wall disappeared, the nucleoli faded and the chromosomes ap- peared. These cells, which a few minutes before had contained only a very few mitochondria, now became full of short dumb- bell-shaped rods, while the resting cells did not undergo any change. So far as could be seen by most careful observation, this increase in quantity of mitochondria was not due to the division of the existing granules. Mitochondria in different kinds of cells Regardless of the fact that the mitochondria constantly change in shape, size and quantity in any one cell, there is a character- istic appearance of the mitochondria in certain kinds of cells, as, for instance, the short, rod and dumbbell shapes are most frequently found in the cells of the endodermal membrane; the long threads, rods and sometimes loops are found more fre- quently in the connective tissue; the small granules and short rods are frequent in nerve fibers and cells, and are often much smaller than those of the connective tissue cells over which the nerve fiber passes; a striated arrangement together with scat- tered granules is characteristic of the fibroblasts; and the large eranules are more frequently seen in the heart and smooth muscle syncitium than in any other kind of tissue. At times the growth from the explanted intestine or heart contains only cells with thread- and rod-shaped mitochondria. Again, a large proportion of such cells contain only large granules. These granules are frequently so large that they are clearly seen with the low power. They are collected about the central body and appear to be more refractive than other types of mitochondria. Occasionally these large granules fuse. That they are not a MITOCHONDRIA IN TISSUE CULTURES oun degenerate form of mitochondria is shown by the fact that such cells frequently divide. In case of mitosis the mitochondria spread around the nucleus, and the large granules become short rods or dumbbell-shaped rods. While these certain characteristic appearances of the mito- chondria are found as a rule in the different kinds of cells, never- theless the shape, position, size and quantity vary so much that it is not always possible to distinguish the kind of cell by the appearance of the mitochondria. [IE 3 / . i] sii ae ae, EE TIE 6 6.00 P.M 6.20 P.M Ammonia vapor Glacial acetic 1% Cc eee e = —_— — ohioy —_—— e = Sa eo ee — CAs et | : 315 PM 4.00 P.M 1.30 P.M. +135 P.M. 1.37-3.00 PM. 3.10 P.M. ‘ 2 qd Normal Hypotonic Sol, Hypertonic Sol, | Hypotonic Sol. Hypertonic Sol. Hypotonic Sol. d Fig. 19 A, effect of 2 per cent glacial acetic acid vapor on the nucleus and adjoining mitochondria and upon a single thread-like mitochondria; b, effect of strong ammonia water vapor on another nucleus with adjoining mitochondria and on a single thread-like mitochondrium, the reaction in both cases was almost instantaneous; c, effect of ammonia vapor on a single mitochondrium followed after 20 minutes by the vapor of 1 per cent glacial zoo acid; d, effect of hypo- and hypertonic solutions on 4 mitochondria. EXPERIMENTAL WORK Mitochondria in the living cell react rapidly and definitely to certain stimuli and in many cases they react more rapidly than either the cell as a whole or any other structure of the cell. This reaction, to be sure, often resembles a disintegration of the mitochondria and results in the rapid formation of varicose mitochondria and then the separation of the varicose mito- chondria into a number of small, finely granular rings. Reaction to acids When the culture is subjected to the action of carbonic acid gas (fig. 13 e) or the vapor of acetic, sulphuric, hydrochloric, chromic and other acids (fig. 19 a, 13 d) the mitochondrial threads rapidly assume a varicose condition and soon separate into a 374 MARGARET R. LEWIS AND WARREN H. LEWIS number of small granular rings of uniform size. Hydrogen per- oxide, potassium permanganate and chlorotone, each produce a similar result. Reaction to alkalies Alkalies, ammonia gas and sodium hydroxide, on the other hand, cause the mitochondria to swell without any sign of vari- cosity. The nucleus also becomes larger and more transparent (fig. 19 b). If the ammonia vapor is followed by vapor from acetic acid the acid will cause the mitochondria and also the nucleus to return to the normal condition. We have not succeeded in stopping the action of the acid at this point, however, and the mitochondria become degenerate rings (fig. 19 ¢). Reaction to xylol, chloroform, ether Xylol, chloroform, and ether simply remove the mitochondrial material, or possibly dissolve the mitochondria and leave shadow forms or slight traces of degenerate mitochondria. Reaction to hyper and hypotonic solutions Changes in osmotic pressure affect the mitochondria often before any change is seen in the cytoplasm. Hypertonic solu- tions shrink the mitochondria while hypotonic solutions cause them to become swollen. The effect of a hypertonic solution can be removed by a decrease in the osmotic pressure of the solution, and, vice versa, that of a hypotonic by an increase in the osmotic pressure (fig. 19 d). Reaction to heat Heat gives interesting results. With an increase in the tem- perature of the warm stage on which the preparation is studied from 40 to 48°C., the mitochondria become round granules within fifteen or twenty minutes, regardless of their previous shape (fig. 20). The size of these round granules is determined by MITOCHONDRIA IN TISSUE CULTURES 375 the size of the mitochondrial thread or rod before the heat began to act. When the heat is applied the mitochondria do not divide into a number of granules, as is sometimes the case when Janus green is used, but each one rounds up as a whole and forms one round granule for each mitochondrium. With rapid cooling of x = > q Rs ~~. —_ = Fi a. ~ <3 vw ae aes eS eae o ‘ ° 6.00 P.M 40°C 600+ PM 46°C 6.01 P.M 47°C. 6.01+ PM. 48°C, 6.02 P.M. 47°C. = _—~ ~> ~ ~~ ="? Pr ee = = — — ° ° ’ x G02 FIM S756 6.03 PM 47°C 6.04PM 46°C 6.05 P.M, 46°C. 6.06 P.M. 45°C. ~) Sa co er on . == =a : > - a ®e Sy ‘ 6.07 P.M 45° : 55 6.08 P.M 45°C. 6.09 PM 45°C 6.10 P.M. 46°C. 6.11 PM. 46°C. We. = a 3a » : : és 3 ’ ‘ ' . = pes) C 612 PM 46°C 613 PM 46°C, 613+ PM. 47°C; “6.14 P.M. 46°C $ Cold water passed = = 1 or 2 migrated &S, through the coils = towards the central 61SPM 46°C of the warm stage iam body; 3 or 4 fol- 618 PM 30°C 6.20. P.M. 44°C lowed at 6.32 P.M = 2 . _— Nw GAS . , as . ed ae cee a > 6.25 P.M 44°C 630 PM 46°C 6.32 P.M. 45°C. 6.35 PM. 48°C. 637/EM 45°C Cold water passed ae > ae = through the coils : f = of the warm stage 645 PM 39°C 647PM 45°C, 7.00 P.M. 48°C. Fig. 20 Part of living cell, drawn with camera lucida, showing position” of 4 mitochondria which were drawn at intervals, while the temperature was first increased and kept at 46 to 48°C. for 15 minutes, then cooled to 39° for 3 minutes and again increased to 44 to 46° for 20 minutes; again cooled to 39° for 8 min- utes, and finally increased to 48° again. a] the preparation, by passing cold water through the coils of the warm stage, the mitochondria return to their normal shape. Prolonged heat, such as 46°C. for over an hour, has in a few instances reduced the number and also the size of the mito- chondria in a given cell. 376 MARGARET R. LEWIS AND WARREN H. LEWIS VITAL DYES Janus green Janus green (di-ethyl saffranin azo di-methyl] aniline) has been considered a more or less specific stain for mitochondria in the living cell, according to Laguesse (’99), Michaelis (99), Bensley (11), Cowdry (12-14). Unfortunately, in our preparations, while the dye stained the mitochondria a brilliant blue-green, it was also toxic to the cells, and even the weakest solution (1-200,000) which definitely stained the mitochondria caused the death of the cells within a few hours. Not only did the dye prevent further growth, but in most instances it also caused various amounts of distortion of the mitochondria. In a few é H ry ‘ a c Fig. 21 Changes exhibited by 4 different cells after application of Janus green; in a and 6b the mitochondria were all long threads before the Janus green was applied and had begun to split up before the drawings could be made. cases the mitochondria moved, changed shape and appeared quite normal, although distinctly stained, but usually the mitochon- drial threads or rods separated into granules (fig. 21), shortly after the stain was applied. This is an indication of a slight degree of degeneration on the part of the mitochondria (see experimental work) and possibly the cell is already injured, although not so greatly as to interfere immediately with the activities of the cell, as in many cases the cell continued to move after the stain had been applied and in one observation on a heart muscle cell in which the mitochondria granules were deeply stained with a (1—100,000) Janus green solution the cell continued to beat for one hour and forty minutes. At the end of this time the stain had faded out and the cell ceased to beat. The dye was dissolved in the Locke’s solution, which was used for that particular explantation, and after a drop of the solution containing the dye warmed to 39°C. had been dropped MITOCHONDRIA IN TISSUE CULTURES 377 on the growth it was drawn off and the tissue again bathed in a fresh drop of the warm solution free from dye. The mito- chondria take up-the dye within a few minutes and remain stained from thirty minutes to two hours. So far as we have observed, the intensity with which the mitochondria stain does not depend upon the strength of the solution. A very weak solution (1—100,000) gives as intensely stained mitochondria as does a strong solution (1-5000). A weak solution, such as 1-100,000 Janus green, stains only the mitochondria a blue- green, while the cytoplasm, nucleus, and nucleolus remain clear. A strong solution (1—-5000), however, stains the cytoplasm a pale green, the mitochondria a darker green, the nucleolus green, and the nucleus a more or less violet-green. Nile blue B extra and brilliant cresyl blue 2b Aside from Janus green, no dye used in these observations stained the mitochondria in the living cell. Both nile blue A concentrated or B extra and brilliant cresyl blue 2 B, however, did stain the mitochondria after the death of the cell, especially after fixation either with neutralized formalin vapor or osmic acid vapor. This is interesting in connection with the work of Lorrain Smith (’08) on differential stains for fats. He states as follows: It was observed that watery solutions of nile blue sulphate (A), a colour stuff of the oxazine series, stains the fat globules contained in tissue cells in various colours. In the majority of cases the fat globules are stained a brilliant red; occasionally globules are present which take a blue stain, and not infrequently the colour is due to a mixture of blue and) red.!° )/"% . We may express the reaction in the following way: The fatty acid combines with the oxazine base to form a blue soap, whereas both neutral fat and fatty acid merely dissolve the rela- tively weak oxazone base (red). He remarks in relation to tissues fixed with formalin that the globules stain readily either red or blue according to their composition: When a globule contains a small amount of fatty acid and a large amount of oxazone base is present in the solution of the dye, the globule becomes predominately red, whereas if the stain is relatively weak in oxazone the blue colour of the oxazine staining is more apparent. 378 MARGARET R. LEWIS AND WARREN H. LEWIS We found that not only do nile blue (A concentrated and B extra) and brilliant cresyl blue (2b) show the above changes of color found by Smith (’08) with fats but also that each dye changes from blue to pink in the presence of certain other sub- stances as shown in table 1. Both brilliant cresyl blue 2b and nile blue B extra are toxic to the cell, and a preparation never lived more than an hour after even the weakest (1—200,000) solution of the nile blue B extra. Brilliant cresyl blue 2 b is less toxic than nile blue and each of these stains is in a way antiseptic, for no infection took place after the stain was used although the dye was not steri- lized. The color reactions with these stains on the living and on the dead cells are shown in table 2. TABLE 1 NILE BLUE | BRILLIANT CRESYL B EXTRA BLUE 2B Sodium carbonate...............; (in solution) | blue _ blue Sodium carbonate............... | (dry) red | bluish violet Lithium carbonate.............. (in solution) | pink | violet Potassium hydroxide............ | pink | pink Sodium hydroxide............... | pink | pink Ammonia Wateiy.c. oa. <3 «sisted 2 precipitate red dirty brown | solution TABLE 2 i me a ch | CYTOPLASM] NUCLEUS |NUCLEOLUS) peoormrs | \(f10.22) | GRANULES | DRIA Living cell Nile blue clear clear clear __refrac- _ pink | blue clear B extra tive | Brilliant clear clear pale | refrac- | pink | purple | clear cresyl blue | blue __ tive | 2b ae oe. Sale oe al ik a: Dead or fixed cell Nile blue | pale blue blue- blue clear | blue | blue B extra blue | violet | | Brilliant | pale _—blue- blue | blue clear purple gray- cresyl blue | violet violet | | | | violet 2b | | & MITOCHONDRIA IN TISSUE CULTURES 379 The difference in the results obtained when these dyes are used upon dead cells and when used upon living cells shows clearly that the chemical conditions which exist in the living cell are quite different from those in the dead cell. What hap- pens in the living cell to prevent the mitochondria and fat glob- ules from taking on the pink or blue color which is assumed immediately upon the death of the cell? Was the dye itself oxidized and why did the vacuoles and certain other granules stain? The vacuole certainly does not take the pink color due to the presence of fat of any kind, for death of the cell would hardly remove the fat but would only change it possibly from neutral to acid fat and the vacuole should then change from pink to blue color instead of fading out entirely. If on the other hand the pink color is due to the alkaline nature of the vacuoles, why then does it not either remain pink or else become blue? Why does the nucleus remain unstained until death of the cell begins and then the nucleolus first take on the stain and later the nucleus? Is the pale blue color of the nucleus after brilliant cresyl blue 2 b in the living cell a delicate indicator that the cell is injured by the dye? These are but a few of the questions suggested by the different action of these dyes upon the dead and the living cell and which must be left for the physiological chemist to solve. This change is most readily seen when a cell has first been stained while it is living and then fixed under the microscope. As the preparation dies the pink vacuoles fade out and the nucleolus, the nucleus, cytoplasm, fat globules and mitochondria stain. This is not due to the direct action of the fixative upon the stain itself since a fixed preparation which has been well washed with Locke’s solution gives the same results with these dyes. Lodine It might be mentioned in this connection that while the vapor from a crystal of iodine did fix the mitochondria as reddish brown threads, rods and granules, there was no evidence of any port wine colored granules of glycogen attached to any 380 MARGARET R. LEWIS AND WARREN H. LEWIS mitochondrium nor within the loop or ring shaped mitochondria. A few glycogen granules were occasionally present, however, as could be distinguished by the color reaction. The fat glob- ules stain first a pale port wine color which later becomes black- ened. If Guilliermond’s (’13) conclusion that the loop-shaped mitochondria give off glycogen granules is correct, one would certainly expect to find that iodine used in connection with un- fixed material would show this at least during the final stage in the formation of the glycogen when the granule lies free but still in the neighborhood of the mitochondrium from which it came. CERTAIN OTHER CELL STRUCTURES AND THEIR RELATION TO THE MITOCHONDRIA Granules Certain other granules were present in most of the cells of these growths, but so far, these granules have not been carefully studied. They can be differentiated from mitochondria of simi- lar shape by the greater rapidity with which the granules move through the cytoplasm. Certain of the vital dyes which color these granules leave the mitochondria unstained. Neutral red usually stains one or several granules near the central body. Nile blue B extra and brilliant cresyl blue 2 b also stain certain granules near the central body. In cells which contain the body we have termed vacuole (see below) one or more of these granules are present within the vacuoles and are stained blue within a pink vacuole (nile blue B extra) or purple within a pink vacuole (brilliant cresyl blue 2b). These granules are few in number in the normal cell but plentiful in cells which contain many vacuoles. Other vital dyes stain certain granules within the cell, but so far as our observations go they are the same as the neutral red granules or else as the nile blue, and brilliant cresyl blue granules. At times the granules within the vacuole take the Janus green color as a very pale green, but no other relation between these granules and the mitochondria has been found. MITOCHONDRIA IN TISSUE CULTURES 381 Vacuoles There are two distinct types of degeneration of the cells of the tissue cultures. The cell either suspends activities, rounds up and dies, or else the cell continues its usual activities but the cytoplasm becomes filled up with vacuoles and the mitochondria become small granules (fig. 22). In a healthy cell a vacuole is often seen to come and go in the cytoplasm, but when several vacuoles remain in the cytoplasm degeneration has begun and the cell never again resumes its normal appearance, but con- tinues to accumulate vacuoles until most of the cytoplasm is used up and only a network which contains scattered granules remains. Fig. 22 Cell from a 3-day culture of intestine from a 7-day chick embryo; the cell has a number of vacuoles near the nucleus, most of the vacuoles contain one or more granules; X 1580 diam. In the fixed and stained preparations the vacuole appears either as a clear space often difficult to differentiate from the fat globule space, or it appears as a clear space within which is a faintly stained granular substance (gray with Heidenhain’s iron hematoxylin or red brown with Bensely’s anilin fuchsin, methylen green stain). In the living cell these vacuoles are distinctly different from the fat droplets. They appear to be fluid spaces not at all refractive, in fact, they resemble a hole in the cytoplasm. Small dancing granules which vary in number from one to many may be suspended in the fluid of the vacuole or closely attached to THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 3 382 MARGARET R. LEWIS AND WARREN H. LEWIS the side. These granules usually stain a pale green with Janus green stain. | Nile blue B extra and brilliant cresyl blue 2b each act as a differential stain for these bodies. The vacuole stains pink and the granules blue (nile blue B extra) or purple (brilliant cresyl ‘blue 2'b, fig. 23): a, ee eee 10.46 A.M. 10.48 A.M. Ay. © © ©) .@554b .@) (Se soa | i Aye 11 A.M. 11.5 A.M. 11.12 A.M. EON 4 ©. wm @ Ge = aes a Ae ani sea 11.30 A.M. b 15pm, 18P.M. \X 11.8 A.M. Soe Q) X11.5 A.M. ®) ca Cc KR d a e . a oss —— @55 o---: oo ~ X12.30 P.M. X12 Noon 12.2 P.M. (8) vac.b. 11.20 A.M. vac.a Vac. b.C © © we ( t a © 2 11.20 A.M. 12 Noon 12.20 P.M. Oo \ [ 4 t ' fg fo 6. ie e) 12.45 P.M. 1.45 P.M. a e ° Vac. c. @) ©) @ Cae e, f° a ! 12.15 P.M. 12.27 P.M. 1P.M. Fig. 23 Observations on the behavior of vacuoles in the living cell with brilliant cresyl blue 2 B (see text). MITOCHONDRIA IN TISSUE CULTURES 383 Brilliant cresyl blue 2b, which is much less toxic than nile blue B extra, shows a most interesting behavior on the part of these vacuoles and granules. A vacuole may appear in the cyto- plasm as a clear unstained space which at first contains no granules, but within five to ten minutes the granules appear as dancing bodies as though they were the result of some conden- sation and precipitation within the vacuole. This process con- tinues and the vacuole takes on first a pale violet color, but later a bright pink, and the granules condense into one or two purple granules. Then the vacuole exhibits various movements such as sending out long pink streamers or threads or becoming U-shaped. Such a vacuole may decrease in size until only the purple gran- ule can be seen. When a cell which contains many vacuoles is stained, all gradations between the pale non-granular vacuole to the single purple granule can be seen. One of the many observations made upon healthy cells in which vacuoles appear is given in full below (fig. 23). 10.45 a.m. 1 gtt. (1-100,000) brilliant cresyl blue 2b in Locke’s solution + 0.25 per cent dextrose was placed on the preparation. 10.50 a.m. The cytoplasm remains clear, several purple granules appear. The mitochondria are unstained, slightly refractive bodies, the fat globules are unstained and highly refractive, the nucleus is unstained, but the nucleolus is a pale blue (fig. 23 a). There is present one vacuole in the cell, which is stained a brilliant pink and contains a large purple granule. The vacuole sends off a long pink streamer quite as long as the thread-like mitochondria but not so thick (fig. 23, vac.a.). The vacuole manifests great activity. The streamer at times becomes detached from the vacuole and fades out. Again it appears to be drawn into the vacuole or sent out from the vacuole. Fig. 23 b. 11.00 a.m. The streamer no longer appears, and the vacuole itself begins to change shape (fig. 23 b) and continued until 11.30 when the vacuole began to grow smaller and a deeper pink. It then remained more or less quiet but grew much smaller in size. 1.05 p.m. The vacuole again sent off a pink streamer which lasted only two or three minutes, after which the small vacuole with one large granule remained quiet. 11.00 a.m. A small granule (x) at the other side of the nucleus moves rapidly to and fro between the nucleus and the periphery several times (fig. 23 c,d, e, f). A streamer of pink follows the purple granule until at 12 noon while the granule moved rapidly towards the nucleus the streamer broke off two pink granules which instantly faded out. « 384 MARGARET R. LEWIS AND WARREN H. LEWIS 12.30 a.m. The granule shows no sign of pink vacuole or streamer and remains quiet near the nucleus. 11.20 a.m. A clear unstained vacuole (vac.b.) appeared in the cyto- plasm between two of the purple granules (fig. 23 g) and behaved as follows, (fig. 23 h). 11.30 a.m. It became a pale violet vacuole with a few dancing un- stained granules. 11.45 a.m. It was a violet vacuole with purple granules. 11.50 a.m. Violet vacuole became pink with purple granules. 12.00 Noon. The vacuole condensed into a small bright pink vacuole with only one purple granule. 12.13 p.m. The vacuole entirely disappeared and only the purple granule remained. 12.20 p.m. Purple granule sent out a pink streamer. 12.45 p.m. The pink streamer osculates and is rapidly sent out and drawn in again. 1.45 p.m. The granule became quiet and the streamer disappeared. The purple granules later moved as a rod from the periphery of the cell in towards the nucleus and back several times. It passed over and under the mitochondria without hindrance. Other purple gran- ules in the cell moved rapidly without streamers, some as double gran- ules, others as rods or as single round granules. 12.10 p.m. A pale space appeared in an adjoining cell (fig. 23 1). 12.15 p.m. This space became pale violet. 12.20 p.m. Violet color changed to pink. 12.25 p.m. Granules appeared in the vacuole. 12.27 p.m. Granules became deep purple granules. The mitochondria and the fat globules remain unstained in all the cells of the growth. In some other cells of the growth many pink vacu- oles are present and also many purple granules. In such cells the mitochondria are mostly small granules and only a few rod- or thread- like ones remain. In the cell under observation the mitochondria did not change type although they were continually changing shape. There was no direct connection between the mitochondria and the formation of the vacuoles in the above observation, and yet in many cells there is often a coincident change in the shape of the mitochondria until in cells which contain many vacuoles within the cytoplasm the mitochondria are no longer in the shape of rods and threads but then appear as small granules. As stated above, the fixed and stained preparations (fig. 24 a) do seem to show all stages in the formation of the vacuoles from the mitochondria just as Dubreuil and Guilliermond have shown the formation of bodies from the mitochondria. How- ever, the fact that the vacuoles have been observed to arise pie MITOCHONDRIA IN TISSUE CULTURES 385 independently of the mitochondria, although there is a coincident change in the shape of the mitochondria makes one exceedingly wary of accepting any evidence from the fixed and stained prep- arations in this regard without corroboration from observations upon the living cell. ——=_- -—>_ —~. a ee d Fig. 24 One-day culture of heart from a 5-day chick embryo; a, cell with various shaped mitochondria similar to those figured by Dubreuil and Guillier- mond, which led them to conclude that mitochondria formed fat and other bodies. Observations on this cell while living gave no evidence for the forma- tion of such bodies from the mitochondria; 6 and ¢ show successive forms of two mitochondria from the above cell and d also shows changes exhibited by a single mitochondrium that Guilliermond might have interpreted as showing the formation of a droplet; e, the rod-shaped mitochondrium which is applied closely to the vacuole was observed, while the cell was living, to migrate from some little distance to the vacuole; it had no connection with the formation of the vacuole. If the specimen had been fixed to show the condition, as in e, one might have concluded that the mitochondrium had something to do with the formation of the vacuole or droplet. Certainly the mitochondria are intimately connected with any change in the cytoplasm, often as in the case of heat without manifestation of change by other bodies in the cytoplasm, and it is probable that any change which takes place in the cyto- plasm such as would cause the formation of vacuoles or other bodies would also have an influence upon the mitochondria of that cell. 386 MARGARET R. LEWIS AND WARREN H. LEWIS Fat globules The connection between the mitochondria and the formation of fat is a very complex and much discussed subject. It has undoubtedly been shown that the mitochondria are bodies which contain lipoid (Fauré-Fremiet ’09; Regaud and Mawas ’09; Fauré-Fremiet, Mayer, Schaeffer, 710; Regaud 710; Mawas ’10; Mayer, Rathery, Schaeffer, °10; Duesberg 711; Dubreuil 713; Cowdry ’14). Our experimental work shows that the mito- chondria act in many ways like bodies which contain lipoid. They are soluble in xylol, chloroform or ether, are slightly black- ened by means of osmic acid, and in fixed preparations are stained blue by means of nile blue B extra and yellow by means of Sudan Ill. It seems probable that the bodies which contain lipoid should form the fat globules, and many observers have tried to establish this (Metzner ’90, Zoja ’91, Loyez ’09, Russo ’09, Dubreuil 713). Others have claimed that the mitochondria are indirectly connected with the formation of fat (Bluntschli ’04, Van der Stricht ’05, Van Durme ’07, Lams and Doorme ’08, Schoonjams ’09). The masterly papers of Dubreuil (11, °13) appear to show clearly and concisely each step in the formation of fat droplets from the mitochondria, and without doubt from the fixed mate- rial which Dubreuil had at hand it seemed to be the logical con- clusion that the fat is formed from the mitochondria. Guillier- mond (’13) in a set of observations equally clear uses many figures similar to those of Dubreuil, but reaches the conclusion that the mitochondria form the glycogen granules of certain cells. It is certainly evident from our observations that no defi- nite conclusions can be drawn from the morphology of the mito- chondria present in any one cell at any one time. Various chemical tests and continued observation of a given mitochon- drium are necessary to establish any morphological conclusion. In our fixed preparations (fig. 24 a) all the figures shown by Dubreuil as evidence that the mitochondria form the fat can be found, i.e., threads, loops, rings and fat droplets, but the study of any one such mitochondrium in the living cell has MITOCHONDRIA IN TISSUE CULTURES 387 never shown that fat droplets arise from mitochondria (fig. 24b, c, d). A thread may form a loop, but the loop changes back again into a thread instead of continuing into a ring. Various rings studied have never changed into globules during observation but have become rods or threads or granules. Such appearances as figure 24 e, were caused by the migration of a mitochondrium to the edge of a vacuole and not as both Dubreuil and Guilliermond might conclude, that the mitochondrium formed the vacuole. Certain granules or thick rods seen in the living cell have the appearance of hollow bodies in the perma- nent preparations and correspond to some of Dubreuil’s figures. This appearance may be due to fixation as Kingsbury (’11) suggests, i.e., that the osmic acid reduced more at the surface and later the more soluble interior is dissolved out. Both Meves (’08) and Duesberg (’11) describe the clear inner part of the mitochondrium, to quote Duesberg, the mitochondria were first present in the early rabbit embryo as small granules but these increase in volume and become large granules at the end of the third day. They have a clear central part with a dark outer edge. Such appearance was seldom seen in the living cell and it is possible that these as well as certain figures of Dubreuil and Guilliermond were formed by the method of fix- ation. Cells which contain both loop and ring shaped mito- chondria frequently show no sign of fat formation, while other cells which are accumulating fat show no mitochondria of the shape which Dubreuil leads us to suppose form the fat droplets. There are three distinct types of fat in these tissue culture growths. First, that in the cells which grow out from tissues that at the time of explantation of the piece of tissue, con- tained fat droplets as the yolk membrane or the migrating fat cells. There seems to be a predetermined ability on the part of these cells to form fat, as is clearly shown where the growth from the yolk membrane adjoins that from the connective tissue (fig. 25). Each new yolk membrane cell contains fat droplets similar to those of the explanted piece of the yolk membrane. In these cells the mitochondria are usually in the form of small granules and the fat droplet is surrounded by granules which 388 MARGARET R. LEWIS AND WARREN H. LEWIS stain like mitochondria. In the migrating fat cells which contain few fat globules some of the mitochondria may be in the form of threads or short rods but there is a coincident change in the shape of the mitochondria with the accumulation of fat drop- lets so that a cell which is crowded full of fat droplets contains Fig. 25 Photograph of part of a 2-day culture of intestine from a 6-day chick. The explanted piece of intestine is from the region where the yolk-sac is attached, and the cells on the left ot the culture are similar to those from cultures of the yolk-sac; each endodermal cell has one or two large fat globules; on the right are mesenchyme cells free or almost free from fat. only small granule shaped mitochondria. The fat droplets are outlined by a row of granules which stain lke mitochondria (fig. 26 b). The second type of fat is one or two small round refractive granules found in almost all the cells of the growths. These fat globules have not been observed to increase markedly in MITOCHONDRIA IN TISSUE CULTURES 389 size or to change their shape. During mitosis they remain stationary and all may pass over to one daughter cell or part to one and part to the other daughter cell. No relation between these fat globules and the mitochondria was observed. Cells which contain one or no fat globules often contain loop or ring shaped mitochondria, but prolonged observation of these has not shown any increase in the amount of fat. The third type of fat is that of an accumulation of fat droplets in many of the Fig. 26 A, b, cells from a 2-day culture of heart from a 10-day chick in which fat is accumulating; a was studied while living, after staining with nile blue B extra, after osmie acid vapor, after nile blue B extra again, Sudan III and Bensley’s aniline fuchsin methylene green. No relation could be found between mitochondria and the formation of fat; b, a typical wandering fat cell with many small granular mitochondria about the fat droplets; c, cell from a 2-day culture of heart from a 11-day chick, accumulating fat; prolonged fixation with osmic acid followed by iron hematoxylin; the fat appears as dark granules. cells of a preparation due to some unknown cause (fig. 26 a, ¢). These cells accumulate fat droplets from day to day, and some cells may become crowded full of fat droplets within forty-eight hours. Such cells should show the relation between the mito- chondria and the fat globules were such a relation present, but so far as our observations go none such could be established. These three types of droplets are undoubtedly fat. It is possible to treat the same cell with various fat stains in succes- 390 MARGARET R. LEWIS AND WARREN H. LEWIS sion and to compare the results. The cell shown in figure 26 a was first studied and drawn while living. The clear, refractive fat globules were easily recognized. A drop of nile blue B extra (1-100,000) was added without any change in the appearance of either the fat globules or of the mitochondria. After a few minutes this was washed off and the preparation was fixed in osmic acid vapor for a few minutes. The same cell was then examined and the fat globules were stained a yellow brown, while the mitochondria remained clear. A drop of nile blue B extra was then added. The fat globules took a dark blue stain and the mitochondria a pale blue. A drop of Sudan III was then added and the fat became yellow while the mitochondria stained a trace of blue. The specimen was then dehydrated and stained with Bensley’s anilin fuchsin, methylene green. The fat droplets were dissolved and the mitochondria stained a bril- liant red. Early in the experimental work it was observed that the mito- chondria under certain conditions became granules around a vesicle. This vesicle stained pink with the nile blue B extra in the living cell, and at that time it was supposed that this indi- cated the formation of a fat by the mitochondria since Lorrain Smith (’08) had shown that nile blue stained neutral fat pink in tissue cells. Later it was observed that nile blue B extra only stains fat in the dead and not in the living cell, and there- fore there was no indication that the mitochondria are in any way connected with the formation of fat. So far as our observations go they show no direct relation between the mitochondria and the formation of fat, although in some cases there is a coincident change in shape of the mito- chondria with the accumulation of fat droplets. Canalicular system One of the interesting cytoplasmic structures, the canalicular system, found by other observers, has as yet not been observed in these living cells. Bensley (11) by means of neutral red observed the small canaliculi as clear spaces in the deeply stained pancreatic cell. MITOCHONDRIA IN TISSUE CULTURES 391 In the cells of tissue cultures neutral red stains only a few gran- ules unless used in such strong solutions as to stain the entire cytoplasm. In such cases a few clear unstained spaces were seen, but a study of the living cell and of the same cell fixed after the neutral red stain by means of osmic vapor and stained with Bensley’s anilin fuchsin, methylen green stain demonstrated that the clear space seen in the cells stained with strong neutral red solution are only the unstained mitochondria. The description of the Binnennetz given by Perroncito (11) certainly resembles in many ways the behavior of the mito- chondria in the tissue culture cells. He finds a network which is like that sometimes seen in these cells, and the ‘corona’ of granules shown in some of his figures appears very much like the mitochondria granules radiating out around the central body. In some of our permanent preparations where vacuoles are present these spaces have all the appearance of the canalicular system. Prolonged fixation in osmic acid did not reveal the canalicular system, although the mitochondria became slightly blackened by the action of the osmic acid. However, none of the special stains for the canalicular system were used, as we desire to deal only with the structures seen in the living cell. Amitosis and giant cells Many cells of these growths contain two or more nuclei and the membrane within the nucleus, which Childs (’07) described as connected with amitosis, is occasionally seen in such cells, but no definite relation between such cells and the mitochondria has been observed. Certainly in some giant cells containing many nuclei, the number of mitochondria present is far greater than that present in a normal cell of the same growth, in fact, it is so much greater that it seems to be definitely related to the amount of nuclear material and to the extent of the cytoplasm. These cells show clearly that there is some other method of increase in the number of mitochondria than that of division at the time of mitosis, for these giant cells appear to be formed 392 MARGARET R. LEWIS AND WARREN H. LEWIS by an amitotic division of the nucleus without a coincident division of the cytoplasm. In regard to the structures of the differentiated cell, such as muscle fibrillae, ete., we have no observations to offer. How- ever, from the behavior of the mitochondria in various shaped cells it is quite evident that any change which affected the morphology of the cell might also change the position of the mitochondria in such a way that they might appear to be con- nected with the formation of the differentiating structure. DISCUSSION We have made no attempt to formulate a theory from the above observations in regard to the origin or function of the mitochondria. A review of the literature shows that the mito- chondria have been found in almost every kind of cell. They are present in the oocyte and spermatocyte (Benda °97, Van der Stricht 00, Meves ’11, and others) and are carried over by the spermatozoon into the egg’ cell in fertilization (Benda ‘11, Meves 711); they are abundant in cells of the young embryo (Meves ’08, Rubaschkin 711); they occur in plant cells as well as in the cells of most animals, including certain of the Protozoa (Lams ’09, Duesberg ’10, Meves ’04, Guilliermond ’12). It is claimed that they form certain cytoplasmic structures such as the fibrillae of the connective tissue (Meves ’10), the neuro- fibrillae in the growing neuroblast (Hoven 710), the myofibrillae (Duesberg 710, Torraca 714) the fibrillae of the epithelial cell (Herxheimer ’89, Korotneff ’09, Fauré-Fremiet 710, Firket 711); that they play a part in the process of cornification (Firket ’11); that they form the secretory granules, directly or indirectly, in the salivary (Regaud and Mawas ’09, Bouin ’05), gastric (Schultze 11), mammary (Hoven ’11) and other glands (Schultze 11). They are described in the rods of the urinary tubule cells (Schultze ’11, Regaud ’08), in the intestinal cells (Champy 10), in the liver cell (Policard ’09). They may form the test of the foraminifera (Fauré-Fremiet ’13). They are described in connection with the formation of the retina cells (Leboueq ’09). MITOCHONDRIA IN TISSUE’ CULTURES 393 Numerous observers have claimed that they form the fat directly (Altmann ’89-’95, Metzner ’90, Zoja 791; Arnold ’07, Russo 07, Loyez ’09, Van der Stricht ’05, Policard ’09, Frissinger 09, Regaud °10, Fauré-Fremiet ’10, Dubreuil 713); indirectly (Bluntschli 04, Van der Stricht 705, Van Durme ’07, Lams and Doorme ’08, Schoonjans ’08). It is claimed that they form the leucoplastids, chloroplastids and chromoplastids and possibly the glycogen (Guilliermond ’12—’13). The above theories seem impossible to correlate. It seems evident that the mitochondria are too universal in all kinds of cells to have the function of forming any one of the above struc- tures of differentiated tissue, and in the light of what cytological chemistry is known, it appears practically impossible for the mitochondria to form all the cell structures mentioned above. In view of the fact that the mitochondria are found not only in almost all animal cells but in plant cells as well it seems more probable that they play a role in the more general physi- ology of the cell. It may be possible that they are concerned with respiration. As suggested by Kingsbury (712), they may represent the structural expression of the reducing substances concerned in cellular respiration, which process Matthews (’05) has described in his theory of protoplasmic respiration. Accord- ing to Matthews, the activity of the cell causes reducing bodies to be formed in the cytoplasm for whose neutralization oxygen is necessary. The hpoid nature of the mitochondria makes it possible to consider them as reducing bodies and certainly the mitochondria exhibit activities which may be due to the fact that they are continually formed in the cytoplasm and con- tinually oxidized. On the other hand, the mitochondria may have to do with assimilation or they may even be stored-up food-stuff themselves, which are continually used up and restored again. Beckwith (’14) holds that the mitochondria are unneces- sary for the life of the cell or for the development of such a complicated structure as a Hydractinia ciliated planula. The fact that such a large group of observers should each have evi- dence to show that the mitochondria form some one structure of the differentiated cell shows that the mitochondria must be 394 MARGARET R. LEWIS AND WARREN H. LEWIS intimately connected with all transformations of the cytoplasm. On the other hand, we must bear in mind the fact that many observers have neglected to identify the body which they had under observation in such a manner that one can be certain that they had the same body which another observer would term mitochondria. It is quite doubtful whether all the bodies called mitochondria are really the same. The criterion for mitochondria in the embryonic cell, as stated by Duesberg after Montgomery, is one which the observer would hesitate to carry out, but some criterion in the sex cell, in the embryonic cells and also in the adult cells should be established for the mitochondria, which all workers will endeavor to fulfil, in order that there may be some common ground for discussion of the results obtained by the numerous observers at work in this field. CONCLUSION 1. Tissue cultures afford an excellent method for observations upon an undisturbed cell as it lives, divides and grows in a medium of known chemical constitution; for experimental work on a living cell; and for the study of the process of fixation. 2. These living cells do not correspond to the usual concep- tion of a cell obtained from the study of fixed material. Both cytoplasm and nucleus are finely granular, almost homogenous in appearance. There is no sign of a reticular or of an alveolar structure of either the cytoplasm or nucleus. Osmic acid vapor is the best fixative for these cells. 3. Mitochondria are present in all the cells of these growths as slightly refractive, large or small granules, rods and threads, similar to those of the chick embryo cell. The mitochondria can be followed and studied in the living unstained cell for hours. 4. The mitochondria may be scattered throughout the cyto- plasm or they may be located around the nucleus or around the idiozome. Any one mitochondrium may change its position in regard to other mitochondria or in regard to the entire cell. Mitochondria located around the centra_some may later migrate MITOCHONDRIA IN TISSUE CULTURES 395 out and become scattered through the cytoplasm, or those scat- tered throughout the cytoplasm may become located around the nucleus. During mitosis the mitochondria become more evenly scattered throughout the cytoplasm, except in the spindle area, where they are usually absent. 5. Any and every shape granule from a minute to a large granule, from small short rods to long threads, loops, rings and networks of various shapes and sizes can be found. Any one type of mitochondria such as a granule, rod or thread may at times change into any other type or may fuse with another mitochondrium, or it may divide into one or several mitochon- dria. Every type of mitochondria is continually changing shape and may assume as many as fifteen or twenty shapes in — ten minutes. The shape of all the mitochondria in a cell can be changed by experimental means such as heat or hyper- or hypotonic solutions. 6. The mitochondria vary greatly in size from minute granules to irregularly shaped, large granules, from short rods to long threads. The size of a single mitochondrium may change by the fusion of two or more granules or by the division of a single mitochondrium. ‘They also appear to increase or decrease without such fusion or division. 7. The number of mitochondria in a single cell varies from two or three to over two hundred. The number of mitochondria is not constant for any one kind of cell or for any phase of any one kind of cell. Daughter cells contain about one-half the number of mitochondria present in the mother cell. The number of mitochondria increases from the daughter cell to the mature dividing cell, and apparently also at times during mitosis. 8. The quantity of mitochondria is not constant for any one kind of cell. Some cells with many small granular mitochondria contain less mitochondrial substance than other cells with a few large granules. 9. Degenerating mitochondria become first a series of gran- ules; later the granules become vesicles and then separate into a number of small finely granular rings which stain like the cyto- plasm rather than like mitochondria. 396 MARGARET R. LEWIS AND WARREN H. LEWIS 10. The mitochondria become more or less scattered through- out the cytoplasm in an indifferent manner and decrease in size during mitosis. About one-half the quantity of mitochondria is separated into each daughter cell by the plane of division. The individual mitochondria pass over entire into one or the other daughter cell and do not each divide into two halves, each going to one daughter cell, as usually described. 11. There are some characteristic differences in the mito- chondria of different kinds of cells, but these are not constant enough to be sufficient to distinguish the kinds of cells. 12. The mitochondria are extremely plastic bodies and often react more rapidly than any other cell structure. They are easily _ influenced in shape and quantity by varous agents, such as heat, carbon dioxide, acids, alkalies, fat solvents, and potassium per- manganate, or by changes in osmotic pressure of the surround- ing medium. 13. The mitochondria are stained in these living cells by Janus green but not by nile blue B extra or brilliant cresyl blue 2 b except in the dead cell. 14. Other granules are present in the cells which are not related to mitochondria. 15. Mitochondria show at times a coincident change in shape with the formation of fat droplets or vacuoles in the cytoplasm, but there is no evidence in these cells of a direct relation between the mitochondria and the formation of either the fat droplets or the vacuoles. 16.. In giant cells the number and quantity of mitochondrial substances is greatly increased above that of the normal cells, somewhat in proportion to the increase in the amount of the cytoplasm and nuclear material. The mitochondria are extremely variable bodies, which are continually moving and changing shape in the cytoplasm. There are no definite types of mitochondria, as any one type may change into another. They appear to arise in the cytoplasm and to be used up by cellular activity. They are, in all probability, bodies connected with the metabolic activity of the cell. MITOCHONDRIA IN TISSUE CULTURES 397 BIBLIOGRAPHY ALTMANN, R. 1894 Die Elementarorganismen und ihre Beziehungen zu den Zellen. Leipzig. ArNoLD, J. 1898 Uber Struktur und Architektur der Zellen. I, 11, III. Arch. f. mikr. Anat., Bd. 52. Barrett, J. O. Wakelin 1913 Changes in the chondriosomes occurring in pathological conditions. Quart. Jour. Micr. Sci., vol. 58. Beckton, H. 1910 On granules in cells of normal tissues and new growths. Arch. Middlesex Hosp., vol. 19. Becxwirtn, C.J. 1914 The genesis of the plasma-structures in the egg of Hy- dractinia echinata. Jour. Morph., vol. 25, no. 2. Brenpa, D. 1899 Weitere Beobachlungen iiber die Mitochondria und ihr Ver. haltnis zu Secretguanulationen, nebst kritischen Bemerkungen. Verh. d. Phys. Ges., Berlin. 1903 Die Mitochondria. Ergeb. d.Anat. u. Entw., Bd. 12. Bensuey, R. R. 1910 On the nature of the canalicular apparatus of animal cells. Biol. Bull., vol. 19, no. 3, p. 192. 1911 Studies on the pancreas of the guinea-pig. Am. Jour. Anat., vol. 12: Buuntscuu, H. 1904 Beobachtungen am Ovarialei der Monascidie Cynthia. microcosmus. Morph. Jahrb., Bd. 32. Bourn, P. 1905 Ergastoplasme et mitochondria dans les cellules glandulaires " _ séreuses. Compt. rend. Soc. Biol., vol. 62. Browne, E. 1913 Study of the male germ cells in Notonecta. Jour. Exp. Zool., vol. 14, no. 1. von Brunn, A. 1884 Samenkérper und ihrer Entwickelung bei Saugetiern und Vogeln. Arch. f. mikr. Anat., Bd. 23. BucHNER, P. 1909-1911 Referate in Arch. f. Zellf., Bd. 2, 5, 6. Crsa, Brancut P. 1910 Contributo alla conoscenza della anatomie e della fisio patologia renale. Internat. Monatschr. f. Anat. u. Phys., Bd. 27. Cuitps, C. M. 1907 Amitosis as a factor in normal and regulatory growth. Anat. Anz., Bd. 30. Craccio, C. 1911 Sui mitochondri degli elementi linfoidi e mieloidi. Patho- logica, vol. 3. Cowpry, E. V. 1912 The relations of mitochondria and other cytoplasmic constituents in spinal ganglion cells of the pigeon. Internat. Monat- schr. f. Anat. u. Phys., Bd. 30. Dusrevin, G. 1911 a Les mitochondries des cellules adipeuses. Compt. rend Soe. Biol. 1911 b Transformation directe des mitochondries et des chondrio- contes en graisse dans les cellules adipeuses. ; 1913 Le chondriome, ete. Arch. d.Anat. Micr., T. 10. DuesserG, J. 1909a Uber Chondrisomen und ihre Verwendung zu Myofib- rillen beim Huhnerembryo. Verh. d. anat. Ges in Giessen. 1909 b Sur la continuité des éléments mitochondriaux des cellules sexuelles et des chondriosomes des cellules embryonnaires. Anat. Anz., Bd. 35, p. 548. THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, NO. 3 398 MARGARET R. LEWIS AND WARREN H. LEWIS DusrsseErG, J. 1910 Les chondriosomes des cellules embryonnaires du Poulet et leur rdle dans la génése des myofibrilles. Arch. f. Zellf., Bd. 4, pp. 602-67. Taf. 28-30, 10 text figs. 1911 Plastosomen, apparato reticolare interno und chromidial ap- parat. Erg. d. Anat. u. Entwick. DuESBERG, J. 1910 Les chondriosomes des cellules embryonnaires du Poulet et leur réle dans la génése des myofibrilles. Arch. f. Zellforschung., Bd. 4. Eisen, G. 1900 The spermatogenesis of Batrachseps. Jour. Morph., vol. 17. Faurt-FREMIET, EK. 1909 Sur les réaction de quelques mitochondries. Compt. rend. Acad. des Sciences de Paris. 1910 Etude sur les mitochondries des Protozoaries et des cellules sexuelles. Arch. d’Anat. micr., T. 11. Faurt-Frimiet, M. E., Mayer, A., AND SCHAEFFER, G. 1910 Sur la micro- chimie des corps gras: Application a l'étude des mitochondries. Arch. d’Anat. micr., T. 12. Frrxet, J. 1911 Recherches sur la génése des fibrilles épidermiques chez le Poulet. Anat. Anz., Bd. 38. GUILLIERMOND, A. 1911 a Surles mitochondries des cellules végétales. Compt. rend. Acad. des Sci., Paris. 1911 b Sur la formation des chloroleucytes aux dépens des mitochon- dries. Compt. rend. Acad. des Sci., Paris. 1913 Sur les mitochondries des champignons. Soc. de Biol. HEIDENHAIN 1911 Plasma und Zelle 2 Lief. ; HERXHEIMER, K. 1889 Uber eigentiimliche Fasern in der Epidermis und Epi- thel verschiedener Schleimhiute. Arch. f. Dermat. u. Syph., Bd. 21. Hoven, H. 1911 Du réle du chondriome dans |’elaboration des produits de sécrétion de la glande mammaire. Anat. Anz., Bd. 39. KoOLLAREWSKY 1887 Inaug. Diso Ber. Kinessury, B. F. 1911 Chromaffin system and the nature of the chromaffin reaction. Anat. Rec., vol. 5. 1912 Cytoplasmic fixation. Anat. Rec., vol. 6. Korscu, Fr. 1902 Die Darstellung des Binnennetzes in spinalen Ganglien zellen und anderen Korperzellen mittelst Osmiumsaiire. Sitzungsber d. k. preuss. Akad. d. Wiss. zu., Berlin. Korotner, A. 1909 Mitochondrien. Chondriomiten und Faserepithel der Tricladen. Arch. mikr. Anat., Bd. 74. 1910 Histologische Beobachtungen iiber die Mitochondrien sowie die Struktur und Entwickelung der Muskelfasern einiger Wirbellosen. Arch. f. Zellf., Bd. 5. LacueEssE, E. 1899 Origine du zymogéne. Compt. rend. Soc. Biol. Lams, H. 1909 Demonstration de mitochondries dans les cellules végétales. Verh. anat. Ges., Berlin. Lams, H., AnD Doormg, J. 1908 Nouvelles recherches sur la maturation et la fécondation de l’oeuf des Mammiféres. Arch. de Biol., T. 23. Lezouca, G. 1909 a Contribution 4 l’étude de l’histogenése de la rétine chez les Mammiféres. Arch. d’Anat. micr., T. 10. 1909 b Etude sur la limitante externe de la rétine. Ann. Soc. Méd. de Gand., T. 89. MITOCHONDRIA IN TISSUE CULTURES 399 | Lewis, W.H., anp Lewis, M.R. 1911 The growth of embryonic chick tissue in artificial media, agar and bouillon. Johns Hopkins Hos. Bull., vol. 22, no. 241, April. 1912 a The cultivation of sympathetic nerves from the intestine of chick embryos in saline solutions. Anat. Rec., vol. 6, no. 1, January. 1912 b The cultivation of chick tissues in media of known chemical constitution. Anat. Rec., vol. 6, no. 5, May. 1912 ¢ Membrane formations from tissues transplanted in artificial media. Anat. Rec., vol. 6, no. 5, May. 1914 Mitochondria in tissue culture. Sci., N. S., vol. 39, no. 1000, February 27. LoyEz, M. 1909 Les premiers stades de la vitellogénése chez quelques Tuni- ciers. Compt. rend. Assoc. Anat., Nancy. 1911 Sur la structure de l’oocyte de la femme 4 la période d’accroiss- © ment. Compt. rend. Assoc. Anat., Paris. Mann, G. 1902 Physiological histology: Method and theory. Oxford, Claren- ~ don Press. Maturews, A. P. 1905 A theory on the nature of protoplasmic respiration and growth. Biol. Bull., vol. 8. Mawas, J. 1910 Etude cytologique et physiologique sur la rétine ciliaire des Mammiféres. Arch. d’Anat. micr., T. 12. Mayer AND RatHEeRY 1909 Recherches sur l’histophysiologie de la sécrétion urinaire chez les Mammiféres. Arch. d’Anat. micr., T. 11. Mayer, RATHERY AND SCHAEFFER 1910 Sur les propriétés des granulations ou mitochondries de la cellule hépatique normale. Compt. rend. Soc. Biol. Merzner, R. 1890 Uber die Beziehungen der Granula zum Fettansatz. Arch. f. Anat. u. Entwick. Meves, F. 1908 Die Chondriosomen als Traiger erblicher Anlagen Cytolo- gische Studien am Hiihnerembryo. Arch. f. mikr. Anat., Bd. 72. 1909 Demonstration von Chrondriokonten in Bindegewebszellen, Knorpelzellen, Osteoblasten und Knochenzellen des Hiithnerembryo. Verh. anat. Ges. Giessen. ‘ 1910 a Zur Einigung zwischen Faden und Granulalehre des Proto- plasma. Beobachtungen an weissen Blutzellen: Arch. f. mikr. Anat., Bd. 75. 1910b Ueber Strukturen in den Zellen des embryonalen Stiitzgewebes, sowie tiber die Entstehung der Bindegewebsfibrillen, insbesondere derjenigen der Sehne. Arch. f. mikr. Anat., Bd. 75, pp. 148-224. 1911 Uber die Beteiligung der Plastochondrien an der Befuchtung des Eies von Ascaris megalocephala. Arch. f. mikr. Anat., Bd. 76. 1918 Uber das Verhalten des plastomatischen Bertandteiles des Spermiums bei der Befruchtung des Eies von Phalluria Manillota. Arch. f. mikr. Anat., Bd. 82, abt. 2, p. 215. Micnaguis, L. 1899 Die vitale Farbung, eine Darstellungsmethode der Zell. granula. Arch. f. mikr. Anat., Bd. 55. Micuaer.is, M., anp Wourr, A. 1902 Uber Granula in Lymphozyten, Vir- chow’s, Arch., Bd. 167. 400 MARGARET R. LEWIS AND WARREN H. LEWIS Misuavsky, A. N. 1911a Demonstration von chondriosomen in Pankreas- zellen. Verh. Anat. Ges., Leipzig. 1911 b Beitrage zur Morphologie der Driisenzelle. Uber das Chon- driom der Pankreaszelle einiger nager. Vorl. Mitteil. Anat. Anz., Bd. 39. Monteomery, T. H. 1911 The spermatogenesis of an Hemipteron, Euschistus. Jour. Morph., vol. 22, no. 3. Perronciro, A. 1911 Beitrige zur Biologie der Zelle. Arch. f. mikr. Anat., Bd. 71: PouicarD, A. 1909 Sur lastructure des mitochondries. Compt.rend. Soc. Biol. 1910 Contribution 4 étude du mécanisme de la sécrétion urinaire. Arch. @’Anat. micr., L. 12. PouicarD, A. Er Garnier, M. 1905 Altérations cadavériques des épithéliums rénaux. Compt. rend. Soe. Biol. 1907 Des lésions rénales provoqués par l’injection de Phlorhizyne. Compt. rend. Soc. Biol. Rawirz, B. 1907 Lehrbuch der mikroscopischen Technik. Leipzig. Ree@aup, Cr. 1909 Participation du chondriome a la formation de grains de séerégation, ete. Compt. rend. Soc. Biol., T. 66. 1910 Etude sur la structure des tubes séminiféres et sur la sper- matogénése chez les Mammiféres. Arch. d’Anat. micr., T. 11. ReEGAwuD, ET Mawas, J. 1909 Sur la structure du protoplasme, ete. Compt. rend. Assoc. Anat., Nancy. Rerzivus,G. 1910 Uber den Bau des Eies derEchinodermen im unbefruchteten und befruchteten Zustand. Biol. Unters. Neue Folge, Bd. 15. RUBASCHKIN, W. 1910 Chondriosomen und Differenzierungsprozesse bei Siuge- tierembryonen. Anat. Hefte, Bd. 41. Russo, A. 1909 I. mitocondri ed i globuli vitellini nell’oocite di coniglia allo stato normale ed in condizione sperimentali. Atti Accad. Gioenia. Catania, Ser. 5, vol. 2. 1910 Sui mutamenti che subiscono i mitocondri ed 1 materiali deuto- plasmici dell’ oocite di coniglia in diversi periodi di inanizione. Arch. f. Zellf, Bd. 4-5. Sakan, Sacucnrt 1913 Uber Mitochondria in die Epidermiszellen der Anuren- larven. Arch. f. mikr., Anat., Bd. 88, Part III. Samssonow, N. 1910 Uber die Beziehungen der Filarmasse Flemmings zu den Faden und Koérnern Altmanns ete. Arch. f.mikr. Anat., Bd. 75. ScHAXEL, J. 1911 Plasma strukturen, Chondriosomen und Chromidien. Anat. Anz., Bd. 39. Scuoonsans, H. 1909 Etude sur la phase d’accroissement des ovocytes chez Ascaris megalocephala bivalens. Bull. Soc. Roy. Sci. Med., Brussels. Scuutrze, O. 1911 Uber die Genese der Granula in den Driiseuzellen. Anat. Anz., Bd. 38. Ssévauy, E. 1906 Uber Spinalganglionzellen usw. Anat. Hefte, vol. 30. Smiru, T. L. 1906 Method for staining fat by means of basic dyes. Jour. Path. and Bact., vol. 11. 1908 On the simultaneous staining of neutral fats and fatty acids by oxazine dyes. Jour. Path. and Bact., vol. 12. MITOCHONDRIA IN TISSUE CULTURES 401 Taxakl, K. 1907 Uber die Stibchenstruktur der Niere. Arch. f. mikr. Anat., Bd. 70. Torraca, L. 1914 Il comportamento dei condriosomi nella regenerazione dei muscoli striati. Arch. f. Zellf., Bd. 12. VAN DER StTRIcHT, O. 1905 La structure de l’oeuf des Mammiféres. II Partie. Bull. de VAcad. Roy. de Belgique. 1905 La structure de l’oeuf de chauve-souris. (V. noctula). Compt. rend. Assoc. Anat., Geneve. Van Durme, M. 1907 Les mitochondries et la méthode de Sjévall dans |’oo- génése des oiseaux. Ann. de la Soc. de Med. de Gand., T. 87. VeJpovsky, F. 1907 Neue Untersuchungen iiber die Reifung und Befruchtung. K6n Bohm Ges. der. Wiss., Prag. Veratti, E. 1909 Sulla fina struttura delle cellule di alcuni timori. Boll. Soc. Med. Chir., Pavia. Wiuson, E. B. 1911 Thecell. Pp. 163. Zora, L BE R. 1891 Intorno ai plastiduli fucsinofili (bioblasti dell’ Altmann). Mem. del. R. Ist. Lomb. di sci., vol. 16. Ai : : rips THE ORIGIN AND EARLY DEVELOPMENT OF THE POSTERIOR LYMPH HEART IN THE CHICK RANDOLPH WEST From the Laboratories of Comparative Anatomy, Princeton and Columbia Universities FOURTEEN FIGURES INTRODUCTION During the autumn of 1912 Professor McClure suggested to the writer the advisability of working out the early develop- ment of the posterior lymph heart in the chick, with especial reference to the source of its endothelium. Throughout the following winter the problem was carried on under Professor MeClure’s supervision at Princeton University, while during the past year it has been continued under the direction of Pro- fessor Huntington at Columbia University. Sala (1) in 1900 described the development of the posterior lymph heart of the bird, and gave a review of the literature to that date. In the caudal sections of an embryo of six days and eighteen hours incubation he finds that: In the mesenchyme which stands in the lateral relation to the caudal myotomes and corresponds to the lateral branches of the first five coccygeal veins, a progressive excavation occurs of little spaces or fissures which soon enter into communication with the lateral venous branches themselvyes—one would say in fact that these fissures are only simple dilatations and ramifications of the veins themselves. If the writer interprets him correctly, Sala states that the lymph hearts are formed by an addition of spaces to the veins, and then a few lines later intimates that these spaces might be considered as “ramifications of the veins themselves.’ He also states that the ‘fissures’ are at first few in number and are arranged in a linear series, parallel to the axis of the vertebral column, corre- 403 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 4 May, 1915 404 RANDOLPH WEST sponding to the point of penetration of each venous branch of the intermuscular septum, and that afterward they gradually increase in number and come to lie near each other. He points out that at the end of the seventh day many of the little ‘fissures’ have fused to give rise to larger spaces, so that the spaces, sepa- rate at first, have finally established irregular communications between themselves, by breaking down their mesenchymal partitions. He goes on to show that by the end of the eighth day the ensemble of the cavities is transformed into a kind of a sac, still communicating with the first five coccygeal veins and later with the general lymphatic system, which develops inde- pendently by fusion of intercellular mesenchymal spaces at first appearing along the veins of the hypogastric plexus. The cavities at this stage often contain red blood cells and some- times appear quite full of them, and by a condensation of mesenchymal cells the wall of the lymph hearts are formed. The rest of this paper, which does not especially concern us, shows that the lymph hearts increase in volume up to the six- teenth day, that the first and fifth coccygeal veins lose their con- nections with the hearts during this period, and that the con- nection of the lymph hearts with the independently developed general lymphatic system occurs toward the end of the tenth day. During the remainder of embryonic life the lymph hearts persist, but shortly after the chick is hatched they commence to degenerate. Traces of the degenerating lymph hearts were found in a chicken thirty-five days after hatching. Mierzejewski, in 1909 (2), published an article on the origin of the lymphatic vessels in birds, which was presented by M. H. Hoyer before the Academy of Sciences of Cracow. Concerning the origin of the posterior lymph hearts he agrees with Sala, except that he holds that the first anlagen appear in the middle of the sixth day of incubation, and not, as Sala states, in the first hours of the seventh day. Stromsten (3) has published two papers in 1910 and 1911 on the development of the posterior lymph heart in turtles. He finds that their development is initiated in the logger-head turtle by the vacuolization of the post-iliaec mesenchymal tissue during ORIGIN OF THE POSTERIOR LYMPH HEART 405 the latter part of the second week of development, and that the spongy tissue thus formed is invaded by capillaries from the dorso-lateral branches of the caudal portion of the postcardinal veins. The capillaries do not communicate primarily with the mesenchymal spaces. Near the close of the third week, parallel veno-lymphatic channels are formed in this spongy area by the confluence of mesenchymal spaces with one another and with the invading capillaries. These veno-lymphatics anastomose freely with each other and communicate by two or three openings with the veins running along their mesial bor- ders. Finally a condensation of mesenchyme and an invasion of muscle cells form the wall, while a confluence of the veno- lymphatic sinuses gives rise to the single sac-like cavity of the adult form of the lymph heart. The subject rested at this point until 1912 when E. L. Clark (4) cited observations, based on injections, to show that in the chick of five days and twenty hours, in the region later occupied by the posterior lymph heart, there exists a lymphatic plexus connected with the coceygeal veins, but not with the haemal capillaries which bear a superficial relation to the lvmphatic vessels. She also shows that the lymphatic plexus is of a differ- ent pattern than the blood capillary plexus and is filled with stag- nant blood, which she considers as backed up from the cocecyg- eal veins. This state of affairs undoubtedly exists in the chick of five days and twenty hours but the observation, aside from its morphological value, throws no light on the origin and mode of growth of the lymphatic plexus. EK. R. and E. L. Clark (5) in a paper in the same number of The Anatomical Record attempt to prove, by observing the first appearance and early growth of this blood filled lymphatic plexus in the living chick embryo of about five days, that it is formed by a purely centrifugal outgrowth from the coccygeal veins. ‘To quote from their article: The first lymphatics in the tail region of the chick arise as direct lateral buds from several of the main dorsal intersegmental coccygeal veins, and not by the transformation of a previously functioning blood vessel plexus. From now on the lymphatic endothelium is 406 RANDOLPH WEST specific and spreads by a steady centrifugal extension The buds send out processes forming clusters. From the clusters, in turn, processes are sent out which anastomose with one another, form- ing a plexus. Simultaneously, processes grow toward the surface from the clusters, and give rise to the superficial plexus of peripheral lym- phaties of the posterior part of the body. There is no essential dif- ference between the manner of growth of the peripheral lymphatics and that of the plexus which is to form the lymph heart (p. 258). In June 1913 Miller (6) in a preliminary note on the develop- ment of the thoracic duct of the chick states that certain aggre- gations of mesenchymal cells mentioned by Sala (1) ‘‘com- prise developing blood cells which are differentiated in situ out of the indifferent mesenchymal syncytium, that these blood cells then gain access to the lymph channelsmaking up the devel- oping thoracic duct, and that finally the haemal cellular ele- ments in question, reach the blood stream via the thoracic duct and the jugular lymph sae.’ He clearly recognizes that lym- phatic channels may serve to transmit blood cells arising 77 situ in the mesenchyme to the haemal channels and distinguishes this function of the lymphatics by the term ‘haemorphic.’ He further states that ‘‘the lymphatics arise as isolated lacunae directly from mesenchymal intercellular spaces and are not in any sense derived from the veins, and subsequently coalesce to form the continuous channel of the thoracic duct.’ The possibility of venous origin of these lymphatics or of the back- ing up of their blood content from the veins is excluded by the total absence of the azygos system in the Sauropsida. In his completed paper of September 1913 Miller (6) gives his results in greater detail. He states that the lacunae in question are bounded at first by indifferent mesenchymal cells which become flattened to form cells which are morphologically equivalent to endothelial cells. Hoyer in June 1913 presented Fedorowiez’s ‘‘Untersuchung iiber die Entwickelung der Lymphgefiisse bei Anurenlarven”’ (7) before the Academy of Sciences of Cracow. Fedorowicz, working on Bufo vulgaris, Bufo viridis, Rana esculenta, and Rana temporaria found cell strands developing from the surface of the lymph heart. In these strands intercellular spaces and Low ORIGIN OF THE POSTERIOR LYMPH HEART 407 finally lumina, which could not be injected from the lymph heart, appeared. The lumen of each strand he found to be lined with endothelial cells. By the continuation of the space formation lymphatic vessels developed which connected second- arily with the similarly acquired lumina of other cell strands which had appeared within the heart. It was not until this connection was established that it was possible to inject the lymphatic vessels from the heart. Allen (8) in a recent important publicatien on Polistotrema (Bdellostoma) describes the caudal lymph heart as arising from isolated mesenchymal spaces in the region of the anterior end of the two branches of the caudal vein, and the ultimate fusion of these spaces by the breaking down of their partitions. Inci- dent to this process certain cells in the interior of the system of spaces become spherical and are transformed into red blood corpuscles. Secondarily the cavity of the lymph heart estab- lishes connections with the caudal vein by the same process, that is a breaking down of mesenchymal partitions while periph- erally the cavity is enlarged by the new formation of isolated mesenchymal spaces and their ultimate annexation. Coinci- dentally the mesenchymal cells bordering the cavity of the lymph heart flatten to form its endothelium. Allen in con- clusion says that his * * * * studies thus far indicate that the most primitive form of lymphatic system are veins that function for both lymphatics and veins. Hence it would be expected that ontogeny would repeat the phylogeny of the lymphatics, and instead of having their origin directly from the veins, they would begin directly as the veins did, by the vacuo- lization of the original mesenchyme. These vessels Allen has designated ‘veno-lymphatics.’ The recognition of haemopoesis in the vicinity of developing lym- phatiecs and from their endothelium is of major morphological importance and the substantial agreement between the results of Allen and of Miller should go far to clear up some of the difficulties that have beset the study of the ontogeny of the lymphatic system. The term veno-lymphatic was used by Huntington and McClure (9) in their studies of the mammalian 408 RANDOLPH WEST jugular lymph sac to designate constituents of the sac which were found at first to contain blood and later to be devoid of blood content. The term veno-lymphatic was simply meant to cover these two conditions of the vessels; for at the time of their studies the criterion of content seemed most available to discriminate between lymphatic and haemal channels. The work of Miller and of Allen demonstrating the 7m situ formation of blood cells and their carriage by lymphatics affords a com- plete and satisfactory explanation of these earlier observations, and Miller’s term haemophoric lymphatic satisfactorily de- scribes the actual conditions, and it is to be hoped in interest of clarity will replace veno-lymphatic. This question was fully considered by Huntington (10) at the Thirtieth Session of the American Association of Anatomists. The present investigation is concerned with the earliest appear- ance of the posterior lymph hearts in the chick. They are two in number and bilaterally symmetrical. Each one arises in the mesenchyme lateral to the caudal muscle plate and posterior to the hind limb bud. Before the lymph heart assumes the form of a single sac-like cavity there exists in this same area a plexus of lymphatic vessels which later coalesce to form the single cavity of the lymph heart. Both the completed lymphatic plexus and later the lymph heart are in connection with several of the most anterior coccygeal veins by means of their lateral branches which pierce the caudal muscle plate, drain the lym- phaties, and then pass outward in the younger embryos to drain a haemal capillary plexus, which bears a superficial relation to the lymphatic plexus. It is the purpose of this paper to show that the plexus of lymphatic vessels, which later enters into the formation of the posterior lymph heart arises by the confiuence of independent mesenchymal spaces which connect secondarily with the veins; that these spaces are bounded at first by mesenchymal cells which become flattened to form an endothelium, and that both in the endothelial lymphatic walls and in the adjacent mesenchyme an active haemopoesis is taking place. ORIGIN OF THE POSTERIOR LYMPH HEART 409 MATERIAL Forty-one of the forty-five embryos used in this work were injected with India ink through the large vitelline blood vessels, the injection usually being pushed to the point of extravasation for the haemal capillaries. Of the four embryos not injected TABLE 1 List of sectioned embryos Length in mm. after Age in days Hours Series fixation Gado 4 12 371 if 4 16 382A 8 4 16 31A 8.5 4 18 5A 8.5 4 18 33A 8.5 4 20 21A 9 4 18 29A 9 4 20 Q27A 9 4 20 23A 9 4 21 22A 9 4 21 4A 9 5 1 7A 9.5 f 18 30A 9.5 4 18 28A 9.5 4 20 244 9.5 4 20 8A 10 1 u 13A 10 +f 21 9A 10.5 4 20 26A 10.5 4 20 25A 11 4 21 20A Wi 5 0 12A 11 5 1 ’ 6A 11 5 13 2A 11.5 5 a 18A LES 5 a 19A eS 5 3 14A 11.5 5 1 10A 12 5 6 34A WH) 5 10 326 13.5 6 1 1A 13.5 te 3A 14 5 20 17A 14.5 ? 11A 15 5 20 15A 15 5 20 16A 410 RANDOLPH WEST through the vitelline vessels three were injected directly into the posterior lymph heart plexus and one (12 mm.) was not in- jected at all. All material was fixed in Zenker’s fluid. Thirty- six of the embryos were cut into 10 » and 7 u serial sections and stained on the slide with eosin and methyl! blue by Mann’s method. One or two series were stained with Delafield’s hemotoxylin and orange G, but this method gave a very poor differentiation of the blood cells. The nine embryos not sectioned were cleared by the Spateholz method and examined in folo under the binocu- lar microscope (table 1). OBSERVATIONS A. FORMATION OF BLOOD CELLS FROM THE MESENCHYME AND THEIR ENTRANCE INTO THE CIRCULATION VIA THE DEVEL- OPING HAEMAL CAPILLARIES, PRIOR TO THE FORMATION OF LYMPHATICS As the appearance of numerous blood cells in the mesenchyme and the extension of the haemal capillaries, previously referred to, is the first change which occurs in the mesenchyme lateral to the caudal muscle plate in the caudal region of the embryo, these processes will be considered first. When the lymphatic anlagen first appear, in the 10.5 mm. embryo, the haemal capil- lary plexus has reached a very high degree of complexity and from this time onward merely holds its own or develops compara- tively slowly. The youngest embryo examined was one of 6.75 mm. In this specimen the mesenchyme lateral to the muscle plate was uniformly loose, and very nearly indifferent. A few rather rounded eosinophile cells were observed in each section. Some of these cells contained one or two large eosinophile granules. Oc- casional venous branches pierced the muscle plate to drain the mesenchyme lateral to it. The same area in the 7 mm. embryo presents several changes. The mesenchyme is much more compact, being equal in den- sity to the mesenchyme which lies medial to the muscle plate. Groups of differentiating blood cells are much more abundant. ORIGIN OF THE POSTERIOR LYMPH HEART 411 These cells are becoming rounded, with a diameter of 7 to 8 u. Their cytoplasm is neutrophile or eosinophile and contains several strongly eosinophile granules. The nucleus is slightly more basophile than the cytoplasm. Eosinophile granules were also observed in the cytoplasm of some of the mesenchyme cells. There is usually a free space of 2 to 3 » about each differentiating cell, which is not encroached upon by the surrounding mesen- chyme. Lateral branches of the coccygeal veins pierce the caudal muscle plate at regular intervals but the capillaries which they drain are few in number. The 8.5 mm. embryo presents a very similar state of affairs, except that the capillaries emptying into the lateral branches of the coceygeal veins are somewhat more numerous, and the differentiating blood cells also occur in greater numbers. As may be seen from figure 1, 5, the haemal capillaries are injected to the point of extravasation, but the differentiating eosinophile cells (7) are absolutely independent of them, nor are there any eosinophile cells medial to the caudal muscle plate. From this stage on until the embryo reaches the length of 10.5 or 11 mm. (fig. 2), the capillary plexus steadily increases in richness and complexity, while the blood cells differentiating from the mesenchyme become scarcer. The capillary plexus has invaded the area formerly occupied by differentiating blood cells, and blood cells in the mesenchyme have decreased until only a small fraction of those present in the 8.5 mm. embryo remap. These blood cells have, then, either degenerated and dis- appeared, or have been drained off by the capillary plexus. The present investigation has not been of such a character as to warrant tracing the complete history of the blood cells which differentiate from the mesenchyme but representatives of both the red and white blood cell lines have been identified in the tissue spaces. That these cells are drained off by the extending capillaries is indicated by the fact that within five or six hours we find first a practically indifferent mesenchyme, a little later a very active haemopoesis taking place in it and finally a general vas- 412 - RANDOLPH WEST cularization of the tissue accompanied by a marked decrease in the number of blood cells in the tissue spaces. It seems highly improbable that decided haemopoesis should take place only to let the cells formed disintegrate three or four hours later without having entered a vessel, and moreover none of the blood cells observed in the tissues appeared to be disintegrating. Mc- Whorter and Whipple (11) in their study of the chick blasto- derm in vitro have observed a to-and-fro movement of the blood cells in the tissue spaces synchronous with the heart beat, and have also observed the entrance of these cells into the general circulation following their rhythmical movement. This phenomenon might be regarded as a plasmatie pulse, which would eventually force any blood cells lying free in the tissue spaces into the general circulation. In addition those cells having the power of amoeboid movement could enter the vessels by diapedesis through the capillary walls. B. DEVELOPMENT OF THE LYMPHATIC PLEXUS AND ACCOM- PANYING HAEMOPOESIS The changes about to be described take place only in the mesenchyme lateral to the caudal muscle plates in the posterior region of the embryo, the mesenchyme lying medial to the muscle plates maintaining its compact indifferent character. For the sake of clearness we shall first consider the Histogenesis and then the Morphogenesis of the developing plexus of lymphatic vessels. Fig. 1 Chick 8.5 mm., Series 21, Slide 1, Row 3, Section 2. X 200. Photo- micrograph of transverse section of caudal end of the embryo. 1, Notochord 5, Haemal capillaries 2, Neural tube 6, Caudal muscle plate 8, Coceygeal vein 7, Differentiating blood cells 4, Coccygeal artery 414 RANDOLPH WEST 1. Histogenesis In the embryo of 10.5 mm. (about 4 days and 22 hours) we observe two new phenomena; the formation of spaces bounded by mesenchymal cells which eventually become flattened to form an endothelium, and the appearance of certain strands of flattened cells in the mesenchyme. Haemopoesis continues to take place in the mesenchyme and also from endothelial cells of the lymphatic walls as soon as these are formed. Throughout the younger stages until the embryo has reached the length of 10.5 mm. the mesenchyme lateral to the caudal muscle plate is of a uniform degree of compactness equal to that of the mesenchyme medial to the muscle plates. The 10.5 mm. embryo, however, shows a slight, but distinct loosening of the mesenchyme just lateral to the muscle plate, between the points of penetration of the lateral branches of coccygeal veins and at certain points the loosening of the tissue is more marked, giving rise to small mesenchymal spaces. The spaces still bounded by mesenchyme are more numerous in the 11 mm. embryo (fig. 2, 8) and some differentiating blood cells have be- come included in them (fig. 5, 7). Certain of the spaces nearest the veins have acquired a venous connection at this stage and in the injected embryos appear as small knob like processes (fig. 5, 10) of a larger caliber than the veins with which they connect. filled with blood cells, and lined by endothelium. These knobs correspond in shape to the mesenchymal spaces men- tioned. It is to be expected that when a space connects witha vein and is subjected to the pressure and friction of the general circulation, that the cells bounding it will tend to become flat- tened. And the fact that in later stages, when the stiil isolated spaces become larger and are under a greater plasmatic pressure Fig. 2 Chick 11 mm., Series 20, Slide 1, Row 4, Section 3. X 200. Photo- micrograph of transverse section of the caudal end of the embryo. 1, Notochord 5, Haemal capillaries 2, Neural tube 6, Caudal muscle plate 3, Coccygeal vein 8, Mesenchymal space 4, Coecygeal artery 9, Lateral branch of coeeygeal vein a i Pi BES EUS ORIGIN OF THE POSTERIOR LYMPH HEART Dot hem +e Pence, > aren Wirt et 10% 2 416 RANDOLPH WEST the bounding cells do flatten, renders it highly probable that a similar process takes place in the case of the smaller spaces which first acquire a venous connection. As was pointed out by E. R. Clark (12) at the Christmas meet- ing of the Anatomical Society in 1913, there are present in the mesenchyme lateral to the caudal muscle plate in the posterior region of the embryo certain strands of flattened cells which Clark holds to be outgrowths from the venous endothelium and to be always capable of being traced back to the veins. These» cells, he says, contain nuclei which may be distinguished from the mesenchyme nuclei by their morphological and staining characters. That strands of flattened cells, sometimes with continuous lumina, sometimes with an interrupted lumen or with no lumen at all occur in the chick as early as 9.5 mm. and more abun- dantly in the later stages, is true. But that they can be clearly distinguished from mesenchyme cells, and that they can always be traced back to a venous endothelium, are at least open questions. E. R. Clark (12) describes the endothelial nucleus as being rather pale and elongated with one or two definite reddish discoid nucleoli, while the mesenchymal nucleus he holds to be darker, and more chromatic with one or two irregular bluish nucleoh, not sharply differentiated from the surrounding chro- matin material. A careful examination, however, reveals a series of graduated stages between these two forms of nuclei. A shght change in the focus of the microscope will make a bluish nucleolus appear reddish, and vice versa, while a careful study of the tissue reveals great variance in the amount of chromatin Fig. 3 Chick 15 mm., Series 16, Slide 2, Row 4, Section 8. X 300. Photo- micrograph of transverse section of the caudal end of the embryo. 1, Notochord 7, Differentiating blood cells 2. Neural tube 8, Mesenchymal space 3, Coceygeal vein 9, Lateral branch of coceygeal vein 4, Coceygeal artery 10, Lymphatic connected with vein 5, Haemal capillaries 11, Aorta 6, Caudal muscle plate Rated: verge § 418 RANDOLPH WEST in the various nuclei. That the typical nucleus of the fully differentiated endothelial cell may be distinguished. from that of the indifferent mesenchyme cell we do not deny, but that in- termediate stages between the two exist, in the case in question we likewise hold to be true. And unless it be cut parallel to its long axis, it is practically impossible to distinguish even the fully differentiated endothelial nucleus from the mesenchymal nucleus. As for the statement that these flattened rows of cells are al- ways connected with a preéxisting endothelium it must be re- membered that practically every cell in the embryo is, at this stage, in syncytial relation with every other cell, the blood cells excepted. So in a certain sense a protoplasmic connection be- tween flattened cells and preéxisting endothelium may be demon- strated by passing over the protoplasm of indifferent mesenchyme cells. To assume that because all endothelium in the embryo is in syncytial relationship it is therefore derived from some preéxisting endothelium, appears unwarranted. Can it not be said with equal truth that since the embryonic vascular endothe- lium is in syncytial relationship with the mesenchyme it is therefore derived from the mesenchyme? This being the case, we know that there are in the mesenchyme certain flattened cells which are not connected with any preéxisting endothelium otherwise than by means of the protoplasm of the mesenchymal syneytium. The isolation of these flattened cells from any other endothelium and the fact that all possible gradations Fig. 4 Chick 8.5 mm., Series 21, Slide 1, Row 3, Section 2. X 500. Fhoto- micrograph of transverse section of the caudal end of the embryo. Fig. 5 Chick 11 mm., Series 20, Slide 1, Row 4, Section 4. X 600. Photo- micrograph of transverse section of the caudal end of the embryo. Figure 4. Figure 5. 3, Coceygeal vein 3, Coceygeal vein 5, Haemal capillaries 5, Haemal capillaries 6, Caudal muscle plate 6, Caudal muscle plate 7, Differentiating blood cells 7, Differentiating blood cells 8, Mesenchymal spaces 10, Lymphatie connected with vein ' ORIGIN OF THE POSTERIOR LYMPH HEART 419 THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 4 420 RANDOLPH WEST exist between them and the typical mesenchymal cells shows clearly that an in situ differentiation of endothelial cells takes place (fig. 6, 13, and fig. 7, 19). The cells so formed may then bound isolated cysts filled with plasma (fig. 6, 72) which some- times enclose a differentiating blood cell. These plasmatocysts then proceed to grow together connecting up with one another and with the veins, and it is probable that they form in some instances a connecting link between the veins and the large lacunae in the mesenchyme. The early appearance of the blood-filled lymphatic plexus connected with the veins in the living chick, which E. R. and E. L. Clark (5) describe as follows, lends weight to such an interpretation of the facts: The first evidence of lymphatics in the tail region of the living chick is the appearance of separate knobs filled with stagnant blood just lateral to the coccygeal veins. Soon after these knobs appear similar ones develop about them which have fne connections with them. * * * * Their injection shows discreet tiny clusters, somewhat like bunches of grapes (p. 254). Figure 6, a section of the caudal region of an 11 mm. embryo, shows an isolated plasmatocyst (7/2). This section and the adja- cent sections were studied with the greatest care under the oil immersion lense, and the two elongated cells (73) with pale nuclei and distinet nucleoli bounding the cyst were not in connection with any other endothelium. Figure 7, a section of the caudal region of a 15 mm. embryo, shows a structure which some might describe as a venous sprout. The injection mass has entered the lumen for a short distance in large amounts. Then the lumen becomes somewhat constricted, and beyond that point only occasional ink granules can be found. Finally the lumen terminates and a long flat cell (73) follows in which two distinet nucleoli are seen, beyond which is a space (75) bounded by a delicate strand of cytoplasm on either side. This space contains a differentiating red blood cell (7). The adja- cent sections have also been examined with great care, and the one directly preceding shows one rather elongated flattened cell with a pale nucleus forming the floor and probably the end of the plasmatocyst containing the blood cell just described. Several ORIGIN OF THE POSTERIOR LYMPH HEART 421 of the mesenchymal cells near by, in the direction in which this ‘sprout’ would extend, show a tendency to become elongated (19), but they are separated from the endothelial cell by indifferent mesenchymal cells, and their nuclei are quite chromatic. They probably represent cells which are about to flatten and to limit a plasmatocyst. Since disconnected plasmatocysts have been found; since all gradations between an indifferent mesenchymal cell and a typical endothelial cell have been observed; and since, in the section just described, we find most distally an uninjected plas- matocyst, containing a differentiating blood cell, then a single endothelial cell enclosing no lumen, and finally a lumen con- nected with the veins into which the injection mass has entered, it does not seem justifiable to call this structure a venous sprout. It should rather be considered as a plasmatocyst which has differentiated in situ, and connected secondarily with the vein. Whether the endothelial cells between the plasmatocyst and the vein arise by an in situ differentiation, or by a mutual growth of the plasmatocyst and the vein toward each other, it is im- possible in this particular case to determine definitely by the study of sections or injections. ‘The latter interpretation would in no way invalidate the fundamental conception that endothelium arises in situ from mesenchyme. It merely implies that en- dothelial cells once formed are capable of proliferation, as cells in general are. It should be noted that discontinuity of the lumen of the ‘sprout’ present in figure 7 shows clearly the utter inadequacy of the injection method for demonstrating all of the endothelium in the embryo. As regards the further development of the blind spaces in the mesenchyme, we have seen that in the 10.5 and 11 mm. embryos there exist a number of spaces in the mesenchyme Just lateral to the caudal muscle plate, and that these spaces are bounded by mesenchymal cells. Some of these spaces are con- nected at this stage with the lateral branches of the coccygeal veins, and certain blood cells, differentiating from the mesen- chyme, have become included in some of the disconnected spaces. In the 12.5 mm. and 13.5 mm. embryos more and 422 RANDOLPH WEST more spaces continue to connect with the veins, either directly or by means of the delicate hollowed ‘cell strands’ already de- scribed, and as the spaces acquire venous connection, they may become filled with blood backed up from the general circulation especially in injected embryos. The spaces which have not as vet attained a venous connection, increase in size, several smaller spaces coalescing by a breaking down of their cell boundaries to form a single larger space (fig. 9, 8; fig. 8, 8; fig. 3, 8). As the plasmatic pressure becomes greater, the indifferent mesen- chyme cells which bounded these spaces become flattened to form cells which are identical in appearance with endothelial cells (fig. 9, 8). The first spaces about which endothelial cells were detected were in a 13.5 mm. embryo, although the cells bounding the spaces were somewhat flattened in the 11.5 mm. and 12.5 mm. embryos. The fact that the cells about a single isolated space may be in part endothelial and in part mesen- chymal, with many intermediate stages between the two, indi- cates that an in situ differentiation of endothelium from mesen- chyme is taking place. The haemopoesis, which was described as taking place before the lymphatic anlagen appear, continues, but much less rapidly than formerly. We have seen that the mesenchyme lateral to the caudal muscle plate was first practically indifferent and non-vascular. Then came a wave of haemopoesis, followed Fig. 6 Chick 11 mm., Series 20, Slide 1, Row 3, Section 7. x 500. Photo- micrograph ot transverse section of caudal end of the embryo. Fig. 7 Chick 15 mm., Series 16, Slide 2, Row 4, Section 6. > 500. Photo- micrograph of transverse section of caudal end of the embryo. Figure 6 7, Differentiating blood cell 8, Mesenchymal space 9, Lateral branch of coceygeal vein ' 13, Elongated cell with pale nucleus and distinct nucleoli 14, Lumen continuous with vein 15, Lumen not continuous with vein 19, Isolated flattened cell, with pale 3, Coceygeal vein nucleus and distinct nucleoli 6, Caudal muscle plate 6, Caudal muscle plate 9, Lateral branch of coccygeal vein 12, Isolated plasmatocyst 13, Elongated cell with pale nucelus and distinct nucleoli Figure 7 424 RANDOLPH WEST quickly by a vascularization of the tissue anda decrease in the number of blood cells in the tissue spaces. ‘This takes the em- bryo up to the 10.5 mm. stage, when the lymphatic anlagen first appear. From this time onward certain mesenchyme cells still seem to become rounded, break away from the surrounding syneytium, and acquire eosinophile granules. In other cells the cytoplasm becomes eosinophile more evenly, forming erythro- cytes. These cells, which le in the tissue spaces, for the most part become included in the lymphatic anlagen, and as these anlagen acquire a venous connection, reach the general circulation. For the first time in the 12.5 mm. embryo groups of rounded strongly basophile cells, may be observed to be differentiating from the endothelium near the junction of the lymphatics and veins. Small clumps of rounded cells, more strongly basophile than the mesenchyme or endothelial cells, are seen forming and apparently splitting off from the endothelium of the lymphatics (fig. 10, 76). In some of the older embryos the cytoplasm of these cells acquires an eosinophile tinge. These cells are identical with the erythroblasts described by Dantschakoff (13). Finally, in the 13.5, 14.5 and 15mm. embryos large aggregations of slightly baso- phile cells with conspicuous eosinophile granules (fig. 11, 17) are seen differentiating and splitting off from the lymphatic endothelium. One final point must be noted, although it does not concern the endothelium of the lymphatic plexus. In the 14.5 mm. embryo strands of three or four myoblasts appear in the now Fig. 8 Chick 14 mm., Series 17, Slide 2, Row 1, Section 6. X 300. Photo- micrograph of transverse section of caudal end of the embryo. Fig. 9 Chick 15 mm., Series 16, Slide 2, Row 4, Section 6. X 600. Photo- micrograph of transverse section of caudal end of the embryo. Figure 8 Figure 9 3, Coccygeal vein 3, Coceygeal vein §, Haemal capillaries 6, Caudal muscle plate 6, Caudal muscle plate 8, Isolated space, bounding cells be- 8, Mesenchymal space coming flattened 10, Lymphatie connected with veins 9, Lateral branch of coceygeal vein 10, Lymphatic connected with vein oe re en pve 5 Oa Pag * 426 RANDOLPH WEST vacuolated mesenchyme just lateral to the caudal muscle plate and parallel to the axis of the notochord, and occasional very small longitudinal spaces may be seen in the most lateral portion of the caudal muscle plate. In one or two sections, one end of the strand of myoblasts was seen to be in connection with the muscle plate. Whether these cells were splitting off from the muscle plate by delamination, or whether they were forming from the mesenchyme and being added to it by accretion, it was not possible to determine in the material available. 2. Morphogenesis Up to this point we have considered the histogenetic changes which take place in the developing lymphatic plexus, and we shall now consider the morphogenesis of the plexus. For this purpose four wax reconstructions have been made by the method of Born, three of which are here reproduced. Chick of 11 mm. Reconstructions of vessels and isolated spaces of the caudal region. X 150. Figure 12: Arteries black, veins and capillaries white, isolated spaces yellow. The posteardinal vein and the aorta run a few sections above the upper level of this reconstruction, but the coccygeal branches of the aorta (fig. 12, 4) and a little more externally the coccygeal veins (3) which drain into the posteardinals, are seen running downward at right angles and dorsal to the axis of the vertebral column. All of these structures are medial to the caudal muscle plate, which has been omitted from this reconstruction for the sake of simplic- ity. This muscle plate extends in a plane, parallel to the ecto- Tig. 10 Chick 12 mm., Series 46, Slide 1, Row 4, Section 4. X 500. Photo-. micrograph of transverse section of caudal end of the embryo. Uninjected. Fig. 11 Chick 15 mm., Series 16, Slide 2, Row 4, Section 7. > 600. Photo- micrograph of transverse section of caudal end of the embryo. Figure 10 Figure 11 3, Coceygeal vein 5, Haemal capillaries 6, Caudal muscle plate 17, Blood cells differentiating from 16, Blood cells differentiating from lymphatie walls endothelium *& ay > a ey 428 RANDOLPH WEST derm, just lateral to the coccygeal veins. ‘Two or three lateral branches of each coccygeal vein (fig. 12, 9) pierce the muscle plate and proceeding directly outward terminate in a plexus of haemal capillaries which lie directly beneath the ectoderm. The lymphatic plexus, which later forms the lymph heart, develops in the mesenchyme between the caudal muscle plate and this superficial plexus of haemal capillaries. A number of isolated spaces, bounded by mesenchyme cells which are still practically unflattened, are seen (fig. 12, 8; fig. 5, 8) to occupy the position just alluded to. They have been studied very carefully with oil immersion lenses and are absolutely inde- pendent of any vascular connection, either with the lateral branches of the coceygeal veins or the haemal capillaries; they occur only caudal to the level of the hind limb bud and only lateral to the muscle plate. Chick of 14 mm. Reconstruction of the blood vessels of the caudal region, and the lymphatic plexus in so far as it forms a continuous channel connected with the veins. X< 150. Figure 13: Arteries black, veins and capillaries white, lymphatics connected with veins, green. The isolated spaces have been omitted from this reconstruction in order that the lymphatic plexus connected with the veins might be more clearly shown. The reconstruc- tion has been drawn from the side and somewhat from above and the aorta and posteardinals have been shown in the draw- ing as folded upward and outward. We have in this recon- struction practically the same arrangement of arteries, veins and haemal capillaries as was described for the 11 mm. embryo. The two posteardinal veins (fig. 13, 78) are seen above and some- what lateral to the aorta; they anastomose above that vessel, and receive the coceygeal veins both cranial and caudal to their anastomosis. The coccygeal veins (fig. 13, 3) as before, pass downward, at right angles to the axis of the vertebral colu nn, close to the caudal muscle plate, and give off lateral branches (fig. 13, 9 a, b) which pierce the muscle plate. It will be seen that a plexus of lymphatic vessels connected with the coccygeal veins has been established between the haemal capillaries and the muscle plate, which is characterized by the irregular size ORIGIN OF THE POSTERIOR LYMPH HEART 429 of its vessels, prominent knob-like enlargements occurring where- ever a large independent space previously existed. This plexus, as has been noted, usually fills with stagnant blood, backed up from the venous circulation. There is no connection between the lymphatic plexus and the haemal circulation except at the point where the lateral branches of the coceygeal veins have just pierced the muscle plate. We now see that the lateral branches of the five or six most cranial cocecygeal veins pierce the muscle plate, drain the lym- phatic plexus and then pass outward to drain the haemal capil- lary plexus (fig. 13, 9a). Soon that portion of the lateral branches of the coccygeal veins distal to the lymphatic taps degenerates, thus severing the connection of these veins with the haemal capillaries, so that those lateral coceygeal branches which drain the lymphatic plexus, cease to function otherwise than for the lymphatic drainage (fig. 13, 9b). An examination of several injected embryos cleared by the method of Spateholz showed this point clearly; the haemal capillary plexus being drained in the 15 mm. embryo by the most dorsal portions of the coceygeal veins with only two of the lateral cocevgeal branches assisting them. although in the embryo of 11.5 mm. five or six lateral coccygeal branches drained the plexus of haemal capillaries. One 17.5 mm. embryo which was examined in cross sections, showed no connection between the lateral branches of the five or six coccygeal veins which drain the lymphatic plexus and the haemal capillaries. Chick of 15 mm. Reconstruction of the caudal vessels. & 150. Antero-lateral view. Figure 14: Arteries black, veins and capil- laries white, lymphatics connected with the veins green, isolated spaces yellow. In this reconstruction the coccygeal veins (3) are seen extending downward from the posteardinals (/8) and the coccygeal arteries (4) from the aorta (//). The coccygeal veins give off lateral branches (9) which pierce the caudal muscle plate—which has been omitted from this reconstruction—and then proceed laterally to drain the lymphatic plexus (green) and at the points where the lymphatics are not as yet formed to any extent, the haemal capillary plexus (white). The lym- 430 RANDOLPH WEST phatic plexus may be clearly seen to occupy the area which in the reconstruction of the 11 mm. chick was filled only by iso- lated mesenchymal spaces. A great number of these isolated spaces (yellow, 8) still exist, not connected as yet with the lym- phatic plexus. They occur in greater numbers medial to the lymphatic plexus which is connected with the veins (green, /0) that is between it and the caudal muscle plate, than they do lateral to the lymphatic plexus, although quite a number, as may be seen from the figure, occupy the latter position. It is especially interesting to note that the isolated spaces le on all sides of the lymphatic plexus, seeming to precede it and form in an area which an hour or two later is occupied by the con- tinuous plexus of lymphatics, connected with the coccygeal veins. Such outlying isolated spaces are clearly shown at the cranial end of this reconstruction. Fig. 12 Reconstruction of caudal vessels of a chick of 11 mm., Series 20. x 150. Antro-lateral view; arteries in black; veins and capillaries in white; isolated spaces in yellow. Fig. 13 Reconstruction of caudal vessels of a chick of 14mm., Series 17. < 150. Antro-lateral view; arteries black; veins and capillaries white; lym- phatie plexus connected with veins, green. The disconnected mesenchymal spaces have been omitted from this reconstruction. Fig. 14 Reconstruction of caudal vessels of a chick of 15 mm. Series 16. < 150. Antro-lateral view; arteries black; veins and capillaries white; lym- phaties connected with veins, green; mesenchymal spaces yellow. Figure 12 9b, Lateral branches of coccygeal vein draining lymphatics only 10, Lymphatic plexus connected with the veins (green) 18, Posteardinal veins 3, Coceygeal vein 4, Coceygeal artery 8, Mesenchymal spaces (green) 9, Lateral branches coccygeal veins draining haemal capillaries Figure 14 Higume ls 8, Coccygeal vein 53, Coccygeal vein 4, Coceygeal artery 4, Coceygeal artery 8, Disconnected mesenchymal space 5, Haemal capillary plexus 9, Lateral branch of coccygeal vein 9, Lateral branches of coccygeal vein 10, Lymphatic plexus connected with 9a, Lateral branches of coccygeal veins veins draining haemal capillaries andlym- 11, Aorta phaties 18, Posteardinal vein ORIGIN OF THE POSTERIOR LYMPH HEART 431 432 RANDOLPH WEST As this investigation has been concerned solely with the origin of the lymphatic plexus which later forms the lymph heart, the later history of the lymph heart has not been studied. ae ee ee ee a a DEVELOPMENT OF THE THYMUS 479 in others they are thinly and quite evenly scattered throughout the basophilic cytoplasm, while still others are completely gorged with them. * A striking peculiarity is that the large majority are mono- nuclear. Only very seldom can one of the polymorphonuclear type be found. The nuclei are round, slightly indented, or crescentic in outline and are usually eccentrically located in the cell. In those cells that are gorged with granules the nuclei are crowded to one edge of the cell and stand out conspicuously among the eosinophile granules. The structure of the nuclei is identical with that of the nuclei in the large lymphocytes which have been described. | The thymus of a 125 mm. embryo was chosen to consider the origin of the eosinophile cells. In this stage the inter- lobular septa, loose in structure, contain numerous lympho- cytes, red blood-cells, and many eosinophile cells lying both singly and in groups. Also in a single group can be found eosinophile cells containing varying numbers of granules, as stated above. An interesting and instructive feature often to be observed is the presence of large lymphocytes containing only from one to three or four eosinophile granules which are of the same size and shape as those found in cells completely gorged with them. Often in very limited areas—covered by very slightly moving the slide under high magnification—can be found large lymphocytes and‘a series of eosinophile cells with gradually increasing numbers of granules (fig. 7, l.m.n.). Only one interpretation can be given to microscopic pictures of this kind, namely, that the eosinophile cells are derived from lympho- cytes. This conclusion also accounts for the large numbers of eosinophile cells along the course of blood vessels in late develop- mental stages, for it is along the blood vessels—in the loose por- tions of the septa—that the lymphocytes are most numerous. I believe that the groups of eosinophile cells in the septa are identical with the granular cells of Watney which he found in the interlobular septa of the thymus in various classes of ani- mals, although none of the cells were attached to the tunica externa of the vessels, as was observed by him. The ultimate 480 J. A. BADERTSCHER source of the eosinophile cells in the interlobular septa of the thymus of the pig is the same as that of the granular cells of Watney, the only difference is that he derived them directly from connective tissue cells while in the pig thymus they are derived from transformed connective tissue cells, the large lymphocytes. Of course, in fixed material it is difficult to determine whether all the lymphocytes along the blood vessels are derived from the loose connective tissue in which the vessels lie or whether some come from the blood. Two features are in favor of the former view; (1) transition forms from connective tissue cells to lymphocytes are of frequent occurrence. The lymphocytes thus formed through division also increase in number; (2) diapedesis of the leucocytes is thought of as taking place only through thin walled blood vessels, but the lymphocytes and eosinophile cells are as numerous along the course of thick walled vessels as along those of a capillary nature. Another possible source of the lymphocytes in the septa is from the par- enchyma of the thymus. However, in late stages that por- tion of the thymus contains mostly small lymphocytes and judging from the small number of small lymphocytes present in the septa very few have migrated into them from the par- enchyma. Only a few eosinophile cells were found undergoing mitosis, so the number of this type of cells formed through their proliferation is almost neglible. The source and nature of all the granules in eosinophile cells is difficult to determine. There is, however, no evidence indi- cating that the granules are débris of degenerated epithelial cells, as held by Schaffer (’09), but ample evidence that not all are products of the protoplasmic activities of the cells containing them, which view is held by Maximow for the ori- gin of the granules of the myelocytes found in the thymus of various animals. Mention was made of free eosinophile gran- ules (fig.,8, Ho.G.) in the interlobular septa where free red blood- cells also occur. These can be traced directly to degenerated red blood-cells, but the free granules usually observed in the septa of any developmental stage do not seem to be numerous | DEVELOPMENT OF THE THYMUS 481 enough to account for all of the granules in the numerous eosino- phile cells even though all should be ingested by lymphocytes. However, lymphocytes with only a few granules in their cyto- plasm and lying among free eosinophile granules suggests that some eosinophile cells are simply lymphocytes ingested with débris of degenerated erythrocytes. This view of the origin of the granules in eosinophile cells is held by Weidenreich (’08, 08, mammals), and by Badertscher (718, amphibia) in a some- what modified form in that some of the granules are also formed from the débris of degenerated muscle tissue. Also circumstan- tial evidence indicating the formation of eosinophile granules from erythrocytes is not wanting and may be enumerated as follows: (1) The free red blood-cells appear in the interlobular septa in advance of eosinophile cells; (2) The red blood-cells appear in large numbers in earlier developmental stages than do large numbers of eosinophile cells, e.g., in the septa of the thymus of a 125 mm. embryo the red blood-cells are more numer- ous than in any other developmental stage while the largest number of eosinophile cells occur in the septa of the thymus of a 165 mm. fetus; (3) As the free red blood-cells in the septa of late stages begin to decrease in number the eosinophile cells decrease in number in correspondingly later stages, e.g., the red blood-cells in the thymic septa of 165 and 180 mm. fetuses are not as numerous as in the 125 mm. embryo but the eosino- phile cells in the 270 mm. embryo are less numerous than in the 165 and 180 mm. fetuses. These facts can be stated in a general way by saying that the height and decrease of erythro- cyte formation in the septa are followed respectively by the height and decrease of eosinophile cell formation in somewhat later stages. If the granules in eosinophile cells are products of degenerated erythrocytes this apparent relationship exist- ing between these two types of cells can be accounted for only on the assumption that the majority of free red cells in the septa undergo dissolution and the products of degeneration taken up by the lymphocytes, possibly in soluble form, and in them transformed into granules. 482 J. A. BADERTSCHER Cells of a peculiar type (fig. 7, &). are quite frequently found among lymphocytes and eosinophile cells in the thymic septa. They are derived from large lymphocytes and are characterized by a part of or the entire superficial layer of the basophilic cyto- plasm staining a deep red similar to the erythrocytes or the granules in eosinophile cells. Their nuclei have the character- istic structure of those in the lymphocytes or eosinophile cells. They cannot, therefore, be erythroblasts which have granular nuclei but must be classed with the eosinophile cells. The cells of this type are never very numerous and the youngest stage in which they. were found was in the body mesenchyme of a25mm.embryo. They occur most frequently in the thymic septa of quite late developmental stages. The origin of the eosinophile cells in the lobules of the thymus can now be discussed briefly. Their structure is the same as of those in the interlobular septa. They belong to the mono- nuclear type. They were first found in the lobules of the thy- mus of a 42 mm. embryo. In this stage they are very rare and can be found only after prolonged searching. Their number increases in successively advanced developmental stages. In the 125 mm. embryo they are readily found in both the cortex and medulla. In the 180 mm. embryo a group of them was found in the medulla of the mid-cervical segment while those lying singly are more numerous than in younger stages. In the full term fetus they are present in appreciably greater num- bers than in the previous stage, groups of them being found in both the cortex and medulla and many can be found lying singly. Since the red blood-cells were considered particularly in the superficial thymus of a 270 mm. (full term) fetus the eosino- phile cells also in that region will be emphasized. Some groups of eosinophile cells are found in the immediate vicinity of blood vessels but as many are found that are not associated with the vessels. The groups occur most frequently along the border of or near the vicinity of groups of erythrocytes but some groups are isolated and as far as position is concerned their origin does not seem to bear any relation to erythrocytes. Here as in the interlobular septa the origin of some is, undoubtedly, from the eee eS Se oe DEVELOPMENT OF THE THYMUS 483 large lymphocytes that have ingested eosinophile granules (débris of degenerated erythrocytes) which as was stated above can be found lying free in the meshes of the reticulum among the lym- phocytes. The free eosinophile granules do not seem to be numerous enough, as in the case of the septa, to account for all granules found in eosinophile cells. However, an apparent general relationship exists between the latter type of cells and the erythrocytes which indicates that at least some of the gran- ules of eosinophile leucocytes are derived from degenerated erythrocytes. The features indicating this relationship may be expressed as follows: (1) As in the thymic septa, the red blood- cells are present in advance of the eosinophile cells; (2) The eosinophile cells increase in numbers in successively advanced developmental stages as do also the red blood-cells; (8) They are most numerous in the thymus of a full term fetus in which developmental stage the red blood cells are also most numerous; (4) In the thymus of a 270 mm. embryo the eosinophile cells are more numerous in the superficial thymus than in the thymus head, the difference in the numbers corresponding favorably to the difference in the numbers of red blood-cells which are much more numerous in the former than in the latter segment. Here also it must be said that if all the granules of the eosino- phile leucocytes in the lobules are derived from degenerated erythrocytes it must also be assumed that their degenerated products are taken up in soluble form by the lymphocytes in which it is transformed into granules. Phagocytes (fig. 7, 0. and p.) are found in the interlobular septa of the thymus in a wide range of developmental stages. They are most numerous in those stages in which the septa have a loose structure and contain many lymphocytes. They possess a large amount of cytoplasm which in some cells is vacuolar. Some are gorged with ingested material which con- sists mainly of lymphocytes (apparently) in various stages of degeneration. Occasionally one can be found in which an entire erythrocyte or a part of one makes up a part of the in- gested material. They arise directly from connective tissue cells some of which contain ingested particles even before their 484 J. A. BADERTSCHER protoplasmic processes have been withdrawn. The phagocytes - vary greatly in size. Some are from two to three times as large as the largest lymphocytes. Only a few were found in the lobules of the thymus of late developmental stages. Cysts were found in the thymus in embryos 55, 65, 110, 125, 165 and 180 mm. in length. They vary in size and shape and all are lined with simple cuboidal or low columnar epithelium which is ciliated only in patches. The cilia are long and slen- der. No consideration was given to their origin. CONCLUSIONS 1. The lymphocytes first present in the thymus are all large lymphocytes and have migrated into it from the mesenchyme. 2. The numerous small round cells of the thymus are formed by the repeated division of the large lymphocytes which thus become small, and also by their own proliferation. 3. Judging from the source and structure of the small round cells they are small lymphocytes and are identical with the small lymphocytes of the blood. The thymus, therefore, may well be considered as a source of some of the small lymphocytes found in the circulating blood. 4. The reticulum of the thymus is of epithelial origin and is formed passively by its meshes becoming filled with lympho- cytes which separate the nodal nuclei farther apart and thus greatly attenuate the protoplasmic processes of the syncytium. 5. The Hassall’s corpuscles are of epithelial origin. 6. The free red blood-cells and eosinophbile cells found in both interlobular septa and the thymic lobules are derived from lym- phocytes in situ. 7. Whether or not any of the erythrocytes formed in the thymus enter the circulating blood is difficult to determine in fixed material. Some of the free erythrocytes undoubtedly undergo degeneration and the products of disintegration of those existing in the form of eosinophile granules are taken up by the lymphocytes which thus become transformed into eosinophile leucocytes. DEVELOPMENT OF THE THYMUS 485 8. It was impossible to trace the origin of all the eosinophile granules in the eosinophile cells directly to degenerated red blood-cells. However, the fact, that the height and decrease of the formation of red blood-cells in the septa is followed by the height and decrease of the formation of eosinophile cells, is cir- cumstantial evidence that a relationship exists between the disappearance of the free erythrocytes and the formation of free eosinophile cells. 9. The histogenesis of the thymus may be divided into epochs each of which is characterized by more or less distinct develop- mental features. They are: (1) The purely epithelial epoch which extends from its origin as an outpocketing from the third pharyngeal pouch and the formation of the cervical vesicle to the appearance of the first lymphocytes in the thymus. (2) The epoch of lymphocyte infiltration and lymphocyte proliferation and the formation of the reticulum. The in- filtration of the thymus by extrathymic lymphocytes from the mesenchyme surrounding it begins in embryos from 25 to 30 mm. in length and probably continues up to stages 180 mm. in length, while their proliferation in the thymus undoubtedly continues after birth. The reticulum, which according to the nature of its development is formed gradually, differentiates ~ into the cortex and the medulla in developmental stages 65 to 75 mm. in length, and is fully formed in embryos 180 mm. in length. (3) The epoch of the formation of red blood-cells and the development of granular cells. An occasional red blood-cell is found in the thymic lobules shortly after lymphocytes are found in them. They are, however, first present in appreciably large numbers in stages of about 55 mm. in length and are most numer- ous in the thymus of full term embryos. In the interlobular septa of the thymus the greatest number occurs in stages of about 125 mm. in length while only a few are found in embryos of 180 mm. in length to full term. EKosinophile cells were first found in the thymic lobules of a 42 mm. embryo but occur first in appreciably large numbers THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 4 486 J. A. BADERTSCHER in embryos of about 180 mm. in length and are most numerous in the parenchyma of the thymus of full term embryos. In the interlobular septa they are seldom found in embryos from 65 to 85 mm. in length. They occur first in appreciably large numbers in the septa of embryos of about 125 mm. in length and are most numerous in embryos 165 to 185 mm. long but are still present in the septa in full term embryos. I wish to thank Prof. B. F. Kingsbury for the aid and encour- agement given me on this work, and Prof. 8. H. Gage for many suggestions. BIBLIOGRAPHY AFANASSIEW, B. 1877 Uber die concentrischen Kérper der Thymus. Arch. f. mikr. Anat., Bd. 14. BapDERTSCHER, J. A. 1913 Muscle degeneration and its relation to the forma- tion of eosinophile leucocytes in amphibia (Salamandra atra). Am. Jour. Anat., vol. 15. 1915 The development of the thymus in the pig. Part I: Morpho- genesis. Am. Jour. Anat., vol. 17. Brearp, J. 1902 The origin and histogenesis of the thymus in Raja batis. Zool. Jahrb. Abt. f. Anat. u. Ontogen., Bd. 17. Betu, E. T. 1906 The development of the thymus. Amer. Jour. Anat., vol. 5. Dustin, A. P. 1911 Le thymus de |’Axolotl. Arch. de Biolog., vol. 26. von Esner, V. 1902 A. Koelliker’s Handbuch der Gewebelehre des Men- schen, Bd. 3, 6 Aufl, p. 328. Goopatu, A. 1905 The post natal changes in the thymus of guinea-pigs and the effect of castration on the thymus structure. Jour. Physiol., vol. 32. Hammar, J. A. 1905 Zur Histogenese und Involution der Thymusdriise. Anat. Anz., Bd. 27. 1908 Zur Kenntnis der Teleostierthymus. Arch. f. mikr. Anat., Bd. 73. 1910 Fiinfzig Jahre Thymusforschung. Kritische Ubersicht der normalen Morphologie. Ergeb. d. Anat. u. Entwickl., Bd. 19. 1911 Zur groberen Morphologie und Morphogenie der Menschen- thymus. Anat. Hefte, vol. 43. Mavrer, F. 1886 Schilddriise und Thymus der Teleostier. Morph. Jahrb., Bde 115 1888 Schilddriise, Thymus und Kiemenreste der Amphibien. Morph. Jahrb., Bd. 13. 1899 Die Schilddriise, Thymus und andere Schlund-spaltenderivate bei der Eidechse. Morph. Jahrb. DEVELOPMENT OF THE THYMUS 487 Maximow, A. 1906 Uber die Zellformen des lockeren Bindegewebes. Arch. f. mikr. Anat., Bd. 67. 1909 a Der Lymphozyt als gemeinsame Stammzelle der verschiedenen Blutelemente in der embryonalen Entwickelung und im postfétalen Leben der Siugetiere. Folia haemat., Bd. 8. 1909 b Untersuchungen iiber Blut und Bindegewebe. II. Uber die Histogenese der Thymus bei Siugetieren. Arch. f. mikr. Anat., Bd. 74. 1912 Untersuchungen iiber Blut und Bindegewebe. IV. Uber die Histogenese der Thymus bei Amphibien. Arch. f. mikr. Anat., Bd. 79. Mirrens, H. 1908 Zur Kenntnis der Thymusreticulum und seiner Bezie- hungen zu dem der Lymphdriisen, nebst einigen Bemerkungen tiber die Winterschlafdriise. Jenaische Zeitschr., Bd. 44. PRENANT, A. 1894 Contribution A l’étude organique et histologique du thy- mus, de la glande thyroide et de la glande carotidienne. La Cellule, e10: Prymak, T. 1902 Beitrige zur Kenntnis des feineren Baues und der Involu- tion der Thymusdriise bei den Teleostiern. Anat. Anz., Bd. 21. ScHaFrrer, J. 1891 Uber das Vorkommen eosinophiler Zellen in der Mensch- lichen Thymus. Central. f. d. med. Wiss. 1893 Uber den feineren Bau der Thymus und deren Beziehung zur Blutbildung. Sitzungsb. d. K. Akad. d.Wiss. Wein., Math.-nat. KI. Abt. 3., Bd. 102. ScHAFFER, J., UND Rast, H. 1909 Das thyreo- thymische System des Maul- wurfes und der Spitzmaus. I. Morphologie und Histologie. Sit- zungsber. d. Wiener Akad., Math. nat. Kl. Abt. 3., vol. 118. Sréur, Pu. 1906 Uber die Natur der Thymuselemente. Anat. Hefte, Bd. 31. Watney, H. 1882 The minute anatomy of the thymus. Phil. Transact. Roya Soc, vole li73;, ps 3: Weipenreicu, F. 1908 a Morphologische und experimentelle Untersuchungen uber Entstehungen und Bedeutung der eosinophile Leucocyten. Anat. Anz., Bd. 32. 1908 b Beitrige zur Kenntniss der granulierten Leucocyten. Arch. i, immer, AMonghes. excl, 7%, REFERENCES ON PLATES A.Erc., amitosis of erythroblasts bl.v., blood vessel D.d.e.N., débris of degenerated epithe- lial nuclei D.e.N., degenerating epithelial nucelus D.e.N’., completely degenerated epi- thelial nucleus E.N., epithelial nucleus Eo.C., eosinophile cells Eo.G., free eosinophile granules Erb., erythroblast Erc., erythrocyte L.L., large lymphocyte L.M., limiting membrane M.C., mesenchymal cell M.e.N., mitosis of epithelial nuclei Me.L., medium sized lymphocyte M.L., mitosis of lymphocyte Nmb., normoblast Pc., phagocyte S.i., interlobular septa S.L., small lymphocyte tial V., vacuole X., cell of unknown origin and _ sig- nificance Z.pr., zone of rapid proliferation of epithelial cells JeIbANIE I) il EXPLANATION OF FIGURES 1 Camera lucida drawing of a portion of a lobule of a section of the right thy- mus head in a 30 mm. embryo. The infiltration of the thymus by extrathymic lymphocytes from the surrounding mesenchyme has just begun. In order to reduce the size of the drawing the very large lymphocyte represented in the outer border of the mesenchyme was drawn a little nearer the thymus than it really is. 1.5 mm. Zeiss App. objective, X 12 Comp. ocular; reduced one-half. 2 Camera lucida drawing of a portion of a lobule of a section of the left thor- acie segment of the thymus in a 37 mm. embryo. The thymus in this develop- mental stage is slightly more advanced in development than in the 30 mm. embryo. This drawing was made to show particularly the large number of epithelial nuclei that are in the first stages of degeneration and the trail that was apparently made by an active lymphocyte that migrated into the thymus from the mesenchyme surrounding it. 1.5mm. Zeiss App. objective, * 12 Comp. ocular; reduced one-half. 488 DEVELOPMENT OF THYMUS IN THE PIG PLATE 1 J. A. BADERTSCHER PLATE 2 EXPLANATION OF FIGURES 3 Camera lucida drawing of portions of two lobules of the left thymus head in an embryo 42 mm. in length. In the thymus of this stage many lympho- cytes are present most of which are large and medium-sized. Only a few small lymphocytes are present. Mitoses cf both epithelial nuclei and lymphocytes occur. The lymphocytes in the interlobular septa are quite numerous. 1.5mm. App. objective, X 4 Comp. ocular; reduced one-fourth. 4 Camera lucida drawing of portions of two lobules of the left thymus head ina 65mm.embryo. The thymus of this stage contains numerous lymphocytes most of which are small ones. Mitosesof lymphocytes are comparatively numer- ous. Many lymphocytes are found in the interlobular septum. 1.5 mm. App. objective, X 4 Comp. ocular; reduced one-fourth. 490 DEVELOPMENT OF THYMUS IN THE PIG PLATE 2 J. A. BADERTSCHER ; =3(¢ oS) gs . 3 & e ae Pee i ¢ aah ey cl ND S55 8) pc PLATE 3 EXPLANATION OF FIGURES 5 Camera lucida drawing of a portion of a thymic lobule in a 36mm. embryo to show specially epithelial nuclei in various stages of degeneration. 1.5 mm. Zeiss App. objective, * 8 Comp. ocular; reduced one-half. 6 Camera lucida drawing of a portion ot a lobule of the thymus in a 55 mm. embryo to show especially free nucleated and non-nucleated red blood-cells. In the portion drawn one erythroblast is in mitotic division while several are in amitotic division. 1.5 mm. Zeiss App. objective X 8 Comp. ocular; reduced one-half. 7 Diagrams showing the different types of cells that are derived from mesen- chymal cells. The direction of the arrows shows the relation of the different types of cells to each other; a, mesenchymal cell; b, c and d, transforming mesenchymal cells; e, large lymphocyte; /, small lymphocyte; g and h, erythroblasts; 7, normo- blast; 7, erythrocyte; k, lymphocyte capped with a layer of hemoglobin; /, m and n show the formation of eosinophile leucocytes, and o and p, phagocytes; a, b, k and p are camera lucida drawings while the remainder are free-hand drawings from actual specimens. All were drawn from specimens in the interlobular septa of a 125 mm. embryo excepting o and p, which were drawn from specimens in an interlobular septum of a 110 mm. embryo. 1.5 mm. App. objective, xX 8 Comp. ocular; reduced one-half. 8 Camera lucida drawing of a portion of an interlobular septum of the thy- mus in a 165 mm. embryo showing specially eosinophile leucocytes and a few free eosinophile granules. 492 PLATE 3 DEVELOPMENT OF THYMUS IN THE PIG J. A. BADERTSCHER 493 ON THE PREMATURE OBLITERATION OF SUTURES IN THE HUMAN SKULL L. BOLK From the Anatomical Institute, University of Amsterdam (Holland) INTRODUCTION In the developmental history of the human skull, there is a period in which the phenomena of development are as yet fairly unknown to us: viz., the phase of life included between the third year and the adult state. The reason of this is quite clear. It is impossible to acquire a perfect knowledge of an object of such intricate structure as the human skull, unless investigation is made of a very great number of infantile skulls. Now, the number of non-adult skulls, except those belonging to children of one and two years old, found in the anatomical institutes is generally quite restricted. This was the case with the anatomi- cal institute of the University of Amsterdam, until two years ago, when it became the possessor of about two thousand infan- tile human skulls. This collection may be utilized for inves- tigations of many totally different natures and my intention is to communicate occasionally in this journal the results of some investigations worked out by myself or by my pupils on the ma- terial of this splendid collection. The present paper will refer to the sutures of the cerebral part of the skull. It is a well known fact, that the bones of the human skull coalesce either during the developmental period or in a more advanced phase of life. In the first case, coalescence takes place for the greater part during the foetal period, in the second case at a date varying extraordinarily for each suture. Therefore in human life a phase exists during which skull-bones do not unite, beginning about the fourth year, when the metopic suture 495 496 ie BOlk has closed itself and the different parts of the occipital bone are united with each other. The sutures still existing at this date of life, are the so-called persisting sutures remaining for a shorter or longer time after the individual has become full grown. Re- garding the variability of age in which these sutures disappear there are already some extensive and carefully written com- munications, for instance by Fredericg and Nibbe. Yet it is a well-known fact that now and then one of these persisting sutures does close during childhood before the individual has reached the adult state. Of this fact, and its influence on the shape of the skull, the casuistic literature is already very abundant, but a systematical inquiry into this phenomenon is as yet wanting. In the present paper I wish to deal with the results of my examination of the premature concrescences of bones in the human skull, results acquired by the investigation of 1820 skulls of non-adult individuals. I regarded the obliteration of the occipito-sphenoidal synchondrosis as a criterion of the adult condition of the skull. The youngest skulls at my disposal already possessed their complete milk-dentition, therefore in this communication skulls of the first two years are not mentioned. Before beginning my investigation I divided my collection into seven groups, in accordance with the developmental phase of the dentition. Of the different groups a brief description follows: Group I. This first group is composed of the very young skulls, with complete milk dentition, and therefore with the following dental formula: ie le Co Wi, TA. These skulls, 725 in number, are those of children who died between the third and sixth year. Group II. This group includes the skulls in which, besides the complete milk dentition, the first permanent molar tooth is also present. The dental formula of these skulls (of children aged 6-7 years) is as follows: i. 1. c. m. m. M. This second set comprises exactly 400 skulls. OBLITERATION OF SUTURES IN SKULL 497 Group IIT. In this group the eruption of the permanent central incisor had taken place. There were 168 of such skulls belonging to children aged about 7 or 8 years. The dental formula is as follows: In tle @5 1m, ittle MME Group IV. This group contains 157 infant skulls in which the lateral permanent incisor has made its appearance corre- sponding with the age of 8 or 9 years. The dental formula of this series is to be written: Ih, Il, @s Wn, 107, IMI. Group V. In this set were included the 109 skulls in which the first milk-molar was lost, and the dental formula is the following: Ile Iho @> 1P5 Tens WH Such a set of teeth corresponds with an age of 9-11 years. Group VI. Includes 203 skulls between 11 and 13 years, in which the second premolar and the canine are present. The order of eruption of both of these teeth is not a constant one. Although in a considerable majority of skulls the second pre- molar precedes the canine, yet there were a certain number in which the eruption of the second premolar evidently succeeds the eruption of the canine. Therefore I have included all these skulls in one set. Its dental formula is as follows: Ialege PME Group VII. This last group contains all the skulls (58 in num- ber) with a complete set of teeth, save the third molar. These skulls, of which the dental formula is PASC. PP. MMe. come from individuals who died between the age of 13 years and the adult state. I have found in my whole collection but three skulls, not yet completely developed, in which the eruption of the third molar had already taken place. It must therefore be considered as a rule that the wisdom tooth makes its eruption after the 498 1. 2B OLK termination of the development of the skull. Exceptions to this rule are very infrequent. Table 1 gives a brief résumé of the above described groups of my collection: TABLE 1 GROUP DENTAL FORMULA AGE NUMBER Lee es ilo He Os fans Toa 3-6 725 lea ee ee ee No Jka (5 Oly wats IE. 6-7 400 Toit, Wyn Pee ee Ils th x 10M foot, IL 7-8 168 1 TRY Geeta em Pee ee 8 [eleven Vie 8-9 157 VE 2 i 2 ceca ees Ws We @5 IRs seals IMI. 9-11 109 Valie Se Sak bo Cone ree ele ee 11-13 203 AVAl Bl ean tre chamen bors Is Wy (Gh, IP. Tes aM, IME 13-20 58 Soon after the beginning of my investigations, the fact struck me that the closing of so-called persisting sutures in skulls of non-adults occurs more frequently than I had supposed. How- ever, this is not the case with all cranial sutures in the same degree. In some a premature concrescence is an unusual rarity, but on the other hand there are some in which the concrescence occurs so often, that it can scarcely be considered as an anom- aly. Now in this communication, I will discuss first: those sutures which I found most frequently closed, and second: those in which coalescence appeared as a very exceptional variety. PREMATURE OBLITERATION OF THE MASTO-OCCIPITAL SUTURE The examination of this suture produced one of the most sur- prising results of my investigation. Fredericg, in his very ex- tensive and valuable paper ‘‘On the obliteration of the cranial suture,’! asserts that the coalescence of the occipital and the temporal bones does not occur before the thirty-first year, it being a very rare exception when it has already occurred in the twenty-first or twenty-fourth year (loc. cit., p. 441). On an- other page in the same work the author strongly points out the 1 Zeitschrift fur Morphologie und Anthropologie, B. 9, 1906. OBLITERATION OF SUTURES IN SKULL 499 fact that the masto-occipital suture belongs to those persisting the longest. Now it is important to note that this author had at his dis- posal only a small number of skulls of 20 to 30 years and that his investigation was made principally on adult skulls. If the investigator had extended his examination upon a sufficient number of non-adult skulls, his conclusion would, no doubt, have been quite different. For the coalescence of the occipital and petrosal bone in the skull of infants is not at all a rare event. On the contrary amongst my material there even was a considerable and unexpected number of skulls, showing complete or partial closure of the masto-occipital suture, either on one side or on both. Moreover not in all cases was the coalescence restricted to this one suture, but in a large number of skulls two or three or even four sutures were totally or partially obliter- ated. Here I wish to treat separately the cases in which only the masto-occipital suture was closed and in which the oblitera- tion was of a more extensive nature. I will begin with the first group. It is scarcely necessary to particularly mention that in case of closure of the masto-occipital suture the coalescence of the two bones can be a total or a partial one. In the second table this fact is taken into consideration. As a rule the coales- cence of the petrosal and occipital bones begins midway in the suture, passing in the majority of cases from this point towards the masto-parietal suture, in consequence of which, in a partial closure, it is most often the upper half which is obliterated. Table 2 shows the results of my examination on the masto- occipital suture. This table demonstrates at once the quite unexpected fact, that in the human non-adult skulls the masto- occipital suture is found closed so often, that one is inclined to consider this phenomenon no longer as an anomaly. Let us consider the frequency of this obliteration. It is not possible to recognize, with the aid of table 2 (p. 500), the absolute num- ber of skulls in which the suture showed obliteration, a certain number being twice mentioned, viz. the skulls entirely closed on the one side and partially on the other; the skulls in which 500 L. BOLK TABLE 2 CLOSURE OF MASTO-OCCIPITAL SUTURE GROUP NUMBER | Both sides Right Left entirely partial entirely partial entirely Partial Tee: Mert SOs Se eG led 12 20 9 21 II 400 12 5 5 11 9 1 | ET Be eee ee 168 10 2 YY 4 3 2 I\ 157 7 1 2 4 2 1 \ 109 6 0) 0 1 2 1 Wiles 203 9 5 8 3 6 5 Walaa ree er 58 3 1 4 1 3 3 25 1820 63 33 44 34 40 on both sides the suture was partially or entirely closed are men- tioned once, and also those in which the suture on one side only was partially obliterated. Moreover an uncertain number re- mains in which the suture is totally closed on one side. Tak- ing this into consideration, I found amongst about 1820 skulls of non-adults at least the number of 63 + 25 + 44 + 40 = 172 with closure of the masto-occipital suture either on both sides or on one side only. Reckoning the number of skulls with a total closure on one side to be 10, we can conclude that in 10 per cent of our non-adults the said suture shows more or less signs of obliteration. Therefore Fredericg’s conclusion is not right, when he writes that the coalescence of the petrosal and occipital bone in the third decennium of life rarely occurs. Even before the twentieth year the coalescence is not exceptional. The preceding table shows yet another phenomenon of no less importance. At what age does this obliteration take place? Our table includes skulls from about three years up to the adult state. Now two possibilities must be considered. Either the coalescence may begin at each date of this period, or the com- mencement of the process is limited to a shorter or longer phase of it. In the first ease the number of synostotic skulls increases while the age advances; in the second case such a correlation is wanting. Now for the solution of this problem it is a happy coincidence that the number of skulls in Group I is considerable. OBLITERATION OF SUTURES IN SKULL 501 This group includes 725 skulls with a complete milk-dentition. And, proceeding in the same manner as before, it appears that in the whole collection the number of skulls showing a closure of the masto-occipital suture in this group must be at least as follows: Complete obliteration on both sides 16 times, partial on both sides 11. A partial closure on the right side only 20 times, and on the left side 21 times, amounting to 68 skulls out of 725. Resuming, we find the following remarkable re- sult. In 1820 skulls varying in age between 3 to 20 years, the masto-occipital suture is obliterated wholly or partially 172 times, making about 10 per cent, and in 725 skulls of infants aged 3 to 6 years, the closure occurred 68 times, also coming to about 10 per cent. In this early stage of childhood, the ob- literation is found in the same proportion as in the total number of skulls including the whole developmental period. Hence the following conclusion is obvious: The number of infantile skulls with closure of the masto-occipital suture does not in- crease after the age of six or seven years; the premature ob- literation of the said suture is limited to a circumscribed period of infancy, beginning as a rule before the end of the sixth year. This fact deserves our full attention in reference to its etio- logical interpretation. For the question arises whether this premature obliteration is a pathological phenomenon, or one of a purely physiological nature. When working out my sta- tistical material I doubted at first the physiological nature. I took into consideration the possibility of this process being caused by some inflammation in the neighborhood of the suture, especially in the tympanic cavity. No doubt an otitis media will cause a hyperaemic state in the adjacent parts of the skull, and one can imagine that under the influence of the latter a coalescence of the occipital and petrosal bone may occur. The consideration, however, that certainly not 10 per cent of our children undergo an inflammation of the middle-ear is sufficient to reject the idea of this pathological cause for the obliteration. Moreover there was yet another circumstance pleading against such a supposition. As we will demonstrate in the following paragraph of this paper, the sagittal suture is also very often the THE AMERICAN JOURNAL OF ANATOMY, VOL. 17, No. 4 502 L. BOLK seat of a premature obliteration, and it is almost improbable that this process is effected by an influence originating from the middle-ear. Therefore it is necessary to explain the great fre- quency of the closure of the masto-occipital suture in infantile skulls in a quite different manner. We will return to this ques- tion after having discussed the premature obliteration of the sagittal suture, which resembles in many points the masto- occipital. The number of all non-adult skulls, showing a closure of the masto-occipital sutures amounts to about 10 per cent, and we have found the same proportion in infants’ skulls aged 3 to 6 years. The process is therefore limited to the period before the commencement of the dentition and is not extended over the whole period of growth of the individual. This fact is proved by another statement given in table 2. It appears that the number of partial coalescences diminishes when the age of the skulls advances and that on the contrary the num- ber of total coalescences increases with the age of the individuals. To prove this let us compare the first two and the last two groups with each other. In the first two groups are included the skulls of children from 3 to 7 years. The total number of these is 1125. In 28, or nearly 25 per cent, of these, a com- plete coalescence of the masto-occipital suture was seen on both sides. In the Groups VI and VII, including the skulls of 12 years and more to the adult state, there were 261 skulls, and of these there were 12, or about 5 per cent, with complete closure on both sides. The difference appears still more considerable by comparing the unilateral coalescence. In the first two groups there are 12 + 5 + 9 + 9 = 35 completely closed sutures on one side and 20 + 11 + 21 +7 = 59 partially obliterated. Therefore, in the very young skulls (Groups I and II) the cases with partial obliteration greatly outnumber those with complete obliteration. After the twelfth year (Groups VI and VII) the correlation becomes reversed; total obliteration being then more common than partial, proved by the following addi- tion: totally closed 8 + 4+ 6+ 3 = 21 and partially closed 3+14+5+4+3 = 12. OBLITERATION OF SUTURES IN SKULL 503 Summarizing the preceding results our investigation leads us to the following conclusions with reference to the masto- occipital suture. In the infantile skull there is found a premature closure of the suture between the mastoid and occipital bone either on one or on both sides in about 10 per cent of the cases. This process is not pathological but ought to be considered as merely physiological. The beginning of the coalescence between the two bones is restricted to earlier stages. After the child has reached its seventh year it has but little chance to be sub- ject to the said premature synostosis. In the preceding pages we only considered the skulls in which exclusively the masto-occipital suture was closed and all others were intact. There were, however, in my collection of infants’ skulls a certain number showing a more complicated condition in which premature obliteration was seen in more than one suture. For the sake of brevity we will postpone the ex- amination of these cases until after the description of the skulls with a single obliteration. THE PREMATURE OBLITERATION OF THE SAGITTAL SUTURE In this suture too my investigation resulted in unexpected results, the frequency of premature closing being more consid- erable than I anticipated. The premature closure of this suture has attracted the atten- tion of many anatomists, more so than the masto-occipital suture. The frequency of the latter’s synostosis was till now an unknown fact in the anatomy of the skull. In general it was acknowledged. that a premature closure of the sagittal suture occurred occasionally, although investigations with statements are as yet wanting. The cause of this difference between two homologous phenomena: is near at hand. In case of closure of the sagittal suture, the possibility arises of a deformity of the skull, more considerable the sooner in life the process com- mences. This anomaly is known as scapho-cephaly (which name was introduced by von Baer), because when excessively deformed the skull becomes boat-shaped. A premature coales- 504 “2, “BOLK cence of the occipital and mastoid bones on the contrary does not cause a striking deformation of the skull or the head. In some cases of a premature union of these two bones I met with a somewhat peculiar form of the occipital region of the skull. But this peculiarity can scarcely be observed during life because the greater part of this region of the skull is covered by the muscles of the neck. Now on the contrary, the deformity be- comes more visible when obliteration of the sagittal suture occurs in early life. The effect of this process is clearly visible and an extensive literature in all the principal languages has treated of this subject. We may distinguish two schools of method in this literature, the purely descriptive and the etio- logical. The investigations of the former simply reported the description of the scaphocephalic skulls, without referring to the origin of the deformity. The naturalists of the latter school did not limit their subject to a simple description, but they went more to the bottom of the problem and tried to point out the genetical cause of the deformity. The opinions of this group were directed in that way principally by a work of Virchow. In it the author dem- onstrates that the anatomical details characteristic of the scaphocephalic skull, were due to the coalescence of the two parietal bones in an early stage of development. Before Vir- chow this hypothesis had been defended by von Baer, but the correlation between cause and effect was clearly demonstrated for the first time by Virchow. However, though I intend to write about the genetical rela- tion between skull deformation and premature obliteration of sutures in a following paper, still I wish to lay stress here upon the justness of an observation already made by Huxley, and which was confirmed by my investigation. This famous natur- alist demonstrated infantile skulls, absolutely normal in shape and size, although the sagittal suture was entirely obliterated. One might observe that in such cases the individual died shortly after the synostosis of the suture and that the skull had there- fore no time to deform. To this supposition I will reply that the number of skulls with premature obliteration and without: OBLITERATION OF SUTURES IN SKULL 505 any sign of scaphocephalic deformation in my collection is too large to accept this point of view. But as mentioned before I will return to this question later on. The number of skulls with premature closure of the sagittal suture was a fairly large one. After finding this fact the ques- tion arose whether this process should be considered either pathological or physiological. To justify the putting of this question some observations may precede upon the variability of the closure of this suture in the adult. The opinions of the writers diverge greatly as to the age in which the normal obliteration of the sagittal suture com- mences. According to Tapmord the process begins normally at the age of 40 to 45 years, a conclusion also accepted by Ribbe. In the text book of human anatomy the average age of the closure is given as about 50 years. Dwight, on the contrary, lays stress upon the fact that the obliteration commences be- tween the twentieth to thirtieth year, although the individual variability is considerable, while the process can occasionally be postponed till a fairly old age. In his admirable paper, already mentioned, Fredericg shows that in 22 out of 34 human skulls, varying between 20 to 30 years, the suture commences to disappear. In this connection the author cited an observation of Schwalbe, who always found the sagittal suture either partially or entirely coalesced after the fortieth year. The process of obliteration however can proceed very slowly, and it even happens that in skulls 80 years of age, the two parietal bones are not yet totally united. Based upon the re- sults of the investigations of Schwalbe and Fredericg, the following point of view presents itself. It is proved, and we need not doubt the reality of the fact, that the beginning of the obliteration of the sagittal suture is seen fairly often between the twentieth and thirtieth year. But this fact was found by merely examining skulls older than 20 years. Until the present time young skulls have not yet been investigated as to the occurrence of the closure of the sagittal suture. And if it becomes clear in the course of such an examination, that such a closure in infantile skulls is not an exception, then I must 506 Li; ‘BOLK say a doubt ought to arise whether such cases have been rightly considered as pathological. It is true that it is premature, for the individual has not attained his adult stage, but why patho- logical? Could it not be possible that the normal variability is even more extensive and that the age at which the obliteration may begin, which as Fredericg truly found, reaches the thresh- old of manhood, may also include a restricted period of childhood? The problem will be thoroughly examined later on. The results of my researches upon the said suture are ar- ranged systematically in the following table 3. TABLE 3 Obliteration of the sagittal suture OBLITERATED GROUP DENTAL FORMULA NUMBER partially entirely Pada’ aviehrntrdes See wae Mla Wy a ceele, sea 725 10 2 1) Se Pees Stic side hy te (ean taal AY 400 8 4 1) Gl May Net At Ae ie a JIS als Gy 10015 soa YE 168 4 2 TLS eA esis Oe Bae Ty @mem:. M 157 3 3 VES Aa a ane ee Tec eee Vi 109 2 2 NARS Nir Peer eee at Al pete: ECE Ve es IML 203 il 6 WALT ey tence sts anaes hicieron seca ee ea Vien iVie 58 0 0 1820 28 19 I wish once more to emphasize that in this table only those skulls. are referred to in which the process of obliteration was limited to the sagittal suture. This table shows that in 47 skulls out of 1820 there was a partial or total closure of the sagittal suture, making 2} per cent. I had not expected to find such a considerable number. The cases of partial obliteration outnumber those of entire closure, a condition which is in no way surprising. For the majority of the skulls are those of children, who died early in life, so that the process of uniting had scarcely time to be extended along the whole suture. In truth the fact that an entire obliteration was found in 19 skulls, making 1 per cent, surprises us as highly as the large OBLITERATION OF SUTURES IN SKULL HOT frequency of the obliteration in general. For by the investiga- tion of Fredericg and Ribbe it is made clear that total oblitera- tion of the sagittal suture in the adult required a fairly long period. Taking this fact into consideration the large number of entirely closed sutures in infantile skulls.awakes a strong sus- picion that the obliteration, beginning in an early period of life, proceeds more quickly than those taking place in the more advanced phase of life. The increased intensity of all physio-+ logical and histological processes natural to youth, evidently influenced also the process of premature obliteration. Now we will enter into the problem whether the obliteration - of the infantile skull is pathological or not. It is clear that this problem is not solved by observing that the union of the two parietal bones, when occurring at an early date in life, causes deformity of the skull to a certain extent, for the effect of an intrinsically normal process may become under circumstances an abnormal one, while the proper nature of the process is not al- tered by it. One must distinguish formal and causal genesis. Moreover one may not conclude that the closure must be of a pathological nature only because it occurs before the full development of the body is reached. For (1) many sutures in the skull disappear during this stage of life and, (2) I call attention to the result of my investigation in which I showed, after examination of about 800 skulls of apes and monkeys,? that in a large number of genera of primates, and especially in anthropoids, synostosis of the sagittal suture happens before the individual is full-grown. Thus, in forms with which the human being stands in close phylogenetical connection, the premature synostosis of the two parietal bones appears to be normal. Here the process bears a purely physiological character. Why should we refuse then to consider it also physiological in man? These arguments however are purely theoretical and through them a decisive answer to the question proposed is not possible. Let us try to find it, by examining more closely the contents of table 3. It showed us that in infantile skulls 2 Zeitschrift fur Morphologie und Anthropologie, B. 15, 1912. 508 L. BOLK the obliteration appeared in 23 per cent. As I pointed out, there are two possibilities. The process is either confined to a definite period of development, or it can happen during its whole course. To determine which of the two possibilities really occurs, we have only to observe the frequency of pre- - mature obliteration appearing in the two groups of youngest skulls, containing those of children from 3 to 6 years. Their total number is 1125. Amongst these skulls there were 24 with partial or entire closure of the sagittal suture, amounting to 2.1 per cent. The conclusion therefore is quite simple and lies close at hand. Amongst 1820 skulls of non-adult individuals (aged 3 to 20 years) there are found 47, or 2.5 per cent, in which the parietal bones are united, and amongst 1125 skulls of children, less than 7 years of age, I count 24, or 2.1 per cent, in which coalescence had taken place. Consequently the number of skulls with synostosis of the sagittal suture scarcely increases after the seventh year. The period during which the obliteration of this suture in in- fancy begins reaches a limit therefore in the seventh year. The tendency to premature closing is not extended over the whole period of growth, but practically stops after the seventh year. I recall the fact that exactly the same relation was found in the masto-occipital suture. Referring to the suture just mentioned, still another cir- cumstance presents itself, proving that the number of prematurely closed sutures do not augment after the seventh year, 7.e., the proportion between *the partially and totally closed sutures. The former diminish as the skulls reach a more advanced age. To demonstrate this I beg the reader to look at the last two rows on table 3. In Groups I, II and III (skulls up to 8 years of age) the partially closed sutures exceed in number those en- tirely closed; in the Groups IV and V (skulls up to 9 and 10 years of age) an equal number of each is found, and finally in Groups VI and VII the entirely obliterated surpass the partially closed ones. I may conclude, therefore, that the process once commenced is of a progressive nature. OBLITERATION OF SUTURES IN SKULL 509 ON THE GENETIC SIGNIFICANCE OF THE PREMATURE OBLITERA- TION OF THE SAGITTAL AND MASTO-OCCIPITAL SUTURES The facts, demonstrated in the foregoing paragraphs as to the sagittal and masto-occipital suture, exhibit so much resem- blance in some principal points, that it is desirable to treat these sutures together from a more general point of view. My reason for intercalating these considerations here and for not waiting till the description of the premature closure of all the sutures is given, is founded on the circumstance that in the other sutures premature obliteration is very seldom seen, and does not occur with the regularity which characterizes the two sutures above mentioned. The following points of resemblance between the two sutures may be mentioned. Firstly, the frequency of pre- mature obliteration. Especially in the masto-occipital suture this is so often found, that one may well question why this phenomenon has remained unknown: in literature until now. The synostosis of the masto-occipital suture is more frequent than that of the sagittal suture. On the other hand one should not forget that the former suture is paired and the chance of a premature closure therefore is doubled. Secondly, both sutures have the fact in common that the commencement of the process of closure is confined to a circumscribed phase of the develop- ment ending approximately in the seventh year. By this limi- tation in time the process attains a peculiar character. There is, aS one might say, in the development of man a stage, dur- ing which he exhibits an intensified tendency to obliteration of some sutures. Beyond this stage, this disposition seems to be lost. The weight of this tendency is not at all a small one, as is proved by the fact that a premature obliteration of the masto-occipital suture is seen in more than 10 per cent of the skulls, and of the sagittal suture in 2.5 per cent. This un- expected large number of cases with premature synostosis gives a predominant significance to the problem of the etiology of this anomaly. This question has already been mentioned, 7.e., is this synostosis of skull bones a pathological phenomenon? In the literature on this subject generally the opinion is ad- vocated that premature synostosis of skull bones is a symptom 510 ; Lie BOLK of some constitutional disease. And as a rule rhachitis or heredity syphilis are accused of being the primary causes of the anatomical anomaly. It is clear that the literature on this subject principally refers to the sagittal suture, because the deformity, which in some cases results from the premature closure of the latter, has long ‘since attracted the attention of anatomists. Concerning this deformity, scaphocephaly as von Baer first called it, an extensive literature exists, in which the question is widely discussed whether scaphocephaly can be acquired after birth, or is in each case already present in the foetal skull. Although we shall not enter into this question, it seems necessary to state the fact that in all the skulls described in the preceding paragraphs, the synostosis of the skull-bones had undoubtedly taken place after birth and in the majority of the cases at the age between the third and seventh year. Still I cannot agree with the opinion of the investigators, who consider the premature obliteration as the result of rhachitical or syphilitical disposition of the individual and will give some arguments against this theory. My first objection is based on the large number of skulls with premature closure. If rhachitis or syphilis is the cause of it, one must not shrink from the conclusion that one or the other of these diseases has affected more than 15 per cent of the individuals. I admit this argument to be purely theoretical and therefore of a problematical value. Still there are other reasons why the pathological nature of the premature obliteration should be denied. In my collection of skulls there were, as need scarcely be mentioned, a certain number with evident symptoms of rhachitis: Hydrocephaly, flattened occipital region, defective development of the enamel of the teeth, ete. A special exami- nation has shown to me that the positive rhachitical skulls were characterized in no manner by an increased tendency to premature synostosis of the skull-bones. Amongst these rha- chitical skulls there were naturally a certain number with pre- mature closure of the sagittal or masto-occipital suture, but the OBLITERATION OF SUTURES IN SKULL Shi proportion in which this happened was not larger than in the wormal skulls. This fact is further strengthened by the cir- cumstance that in most cases of premature obliteration no other symptoms of rhachitis were visible, they possessed a normal structure of the bone tissue and of the tooth-enamel. Another argument pleading against the rhachitical character of the premature obliteration is the great regularity with which the process commences and proceeds. In all the skulls de- scribed in the foregoing paragraph, it was clear that the synos- tosis of the sagittal suture regularly commenced at the very point where in normal cases the obliteration begins, 7.e., in the obelion. Should the process be of a pathological nature the starting-point of the synostosis should be very inconstant. Finally, if the obliteration is really the effect of some general constitutional disease, how can we understand that the process confines itself to the whole length of one suture only? In the sagittal suture the synostosis is often complete, extending from the bregma to the lambda point. Why, one may ask, does not the process continue along the coronal and lambda sutures? Is such an anatomically strictly confined extension of the proc- ess in accordance with a supposed pathological origin? I must admit that these arguments prove nowhere decisively that the premature synostosis cannot be caused by some con- stitutional disease. But on the whole I think that they form a strong evidence against it. On the other hand, I will by no means absolutely deny all genetical correlation between anom- alies in the system of sutures of the infantile skulls and con- stitutional diseases. I willingly admit the possibility of such a relation, but I wish to reserve it for those cases in which an en- tire or partial closure of several sutures is seen in an often very irregular manner. ; Now the question arises as to the real significance of the pre- mature closure. If it is not, as I just made clear, the result of some pathological cause, from which point of view is the phe- nomenon to be explained? I believe I am able to give such an explanation, and I wish to give in the following pages a brief account of my opinion upon this subject. ore Lb. BOLK Some years ago I published an extensive investigation upon the normal obliteration of the sutures in Primates. The re sults of this inquiry were based upon the examination of more than 800 skulls of platyrrhine and catarrhine monkeys and a considerable number of skulls of anthropoid apes, all present in the anatomical museum of the University of Amsterdam. As to the problem interesting us in the present paper, we may limit ourselves to the conclusions relative to the anthropoid skulls. There are striking differences in the process of obliteration between man and apes. These differences concern the age in which normal obliteration takes place and the order of suc- cession in which the closure in the different sutures begins. In man, as a rule, the principal sutures persist for a longer or shorter time after the complete formation of the skull. The same happens in some genera of American monkeys; but in apes the sutures can close immediately after the skull is full grown. At this moment the general growth of the individual is not yet finished, and though it is, for reasons near at hand, impossible to know the age in which the obliteration begins, it is sure that the process commences, and perhaps in some sutures is even finished, before the animal has attained its adult state. The significance of this premature synostosis of the skull bones in apes may be found in the strong development of the muscles of mastication, arising from almost the whole surface of the braincase, and moreover in Gorilla and male Orangs from strong crests developing exactly in the line of union of the parietal and occipital bones. Now it is obvious that in the apes, as well as in man, there exists: a relation between the growth of the brain and the brain- case. In apes, as a rule, the different bones of the skull can- not unite together before the brain has attained its final volume. This is so clear and simple that it is altogether unnecessary to enlarge upon it. As it is, the conclusion lies close at hand that the sutures in the braincase of apes disappear immediately after their physiological function is finished. The physiological s OBLITERATION OF SUTURES IN SKULL 513 function of the sutures is to produce new osseous tissue along the margin of the skull bones for the sake of the enlarging of the braincase. This function is continued as long as the brain- case needs enlarging, 7.e., as long as the brain increases in volume. Summarizing, I think it is clear, that in apes the sutures com- mence to obliterate as soon as the enlargement of the brain has ceased. And in this respect there is a remarkable differ- ence between man and apes. In the former the sutures often persist a long time after the brain has ceased growing. We can now return to our starting point and consider the question whether there is some relation between the normal progress of suture-obliteration in apes and premature oblitera- tion in man. There is no doubt about the fact that man stands in nearer phylogenetical relation to the anthropoids than to any other representative of the primate stem. Therefore, since, as a rule, the sutures begin to disappear in apes shortly after the brain is full-grown, which happens in youthful animals, we have the right to conclude that the condition in man is of a progressive nature. This condition, 7.e., the persistence of the sutures during a certain period of the adult state, must be considered as a peculiarity acquired by man during the earliest phase of his phylogenetic evolution. This conclusion gives rise to the following question. Should not the premature obliteration of the sutures in the braincase of man be considered an atavistic phenomenon? ‘This hy- pothesis deserves our full attention. If the statement is ac- cepted as true, that in human ancestors the sutures closed as those of the anthropoids of today, 7.e., at an early stage of life, then the occasional premature obliteration in recent man loses its non-proved pathological character and becomes more in- telligible. For we know that each quality newly acquired in the evolution of beings often requires a long space of time be- fore it becomes absolutely fixed. During this period the ante- cedent condition reappears individually now and then. For my part I think I may conclude that the premature closure of sutures in infant-skulls is such an atavistic phenomenon. 514 BOLE ON THE OBLITERATION OF ONE OF THE OTHER SUTURES OF THE SKULL In considering the occurrence of premature obliteration a striking difference is observed between the sagittal and masto- occipital suture on the one hand and all the other sutures on the other.